CN114564749B

CN114564749B - User information protection method and server for smart cloud service

Info

Publication number: CN114564749B
Application number: CN202210207155.7A
Authority: CN
Inventors: 黄少龙; 刘厚品
Original assignee: Xiamen Herocheer Electronic Technology Co ltd
Current assignee: Xiamen Herocheer Electronic Technology Co ltd
Priority date: 2022-03-04
Filing date: 2022-03-04
Publication date: 2022-12-23
Anticipated expiration: 2042-03-04
Also published as: CN114564749A

Abstract

The invention provides a user information protection method and a server for smart cloud service, wherein information anonymization operation is performed on individual privacy, and attention indexes (such as characteristic dimensionality) of privacy description mining results of continuous cloud service user interaction records can be considered in the privacy anonymization processing process, so that anonymization processing of relatively important user information can be realized as accurately as possible, anonymous processing of a large amount of user information is not required, the expenditure of system resources is reduced, and the user information protection efficiency is improved.

Description

User information protection method and server for smart cloud service

Technical Field

The invention relates to the technical field of cloud services, in particular to a user information protection method and a server for smart cloud services.

Background

Cloud services are an augmentation, usage, and interaction model for internet-based related services, typically involving the provision of dynamically scalable and often virtualized resources over the internet. The cloud service is to obtain a desired service in an on-demand, easily extensible manner through a network, and is light-asset.

With the continuous development of cloud services, the information security problem of users is also deeply concerned by users when the services are performed in an on-demand and easily-extensible manner through a network. Based on this, the inventors have long and deeply studied and found that a protection method for related user information generally performs uniform anonymous protection processing on a large amount of user privacy information, which not only causes resource waste, but also reduces the protection efficiency of the user privacy information.

Disclosure of Invention

The invention provides a user information protection method and a server for smart cloud service, and adopts the following technical scheme in order to achieve the technical purpose.

The first aspect is a user information protection method for smart cloud service, which is applied to a smart cloud server, and the method at least includes: determining a privacy description mining result of a service session behavior event of a cloud service user interaction record to be subjected to user information protection according to a cloud service user activity log; updating the privacy description mining result based on the attention index, and performing personalized privacy description mining and information anonymity operation on the cloud service user interaction record by using the privacy description mining result of which the description size is updated.

The information anonymization operation is carried out aiming at individual privacy, and attention indexes (such as characteristic dimensions) of privacy description mining results of continuous cloud service user interaction records can be considered in the privacy anonymization processing process, so that anonymization processing of relatively important user information can be accurately realized as far as possible, anonymization processing of a large amount of user information is not required, the expenditure of system resources is reduced, and the efficiency of user information protection is improved.

For some design ideas which can be independently implemented, determining a privacy description mining result of a service session behavior event of a cloud service user interaction record to be subjected to user information protection according to a cloud service user activity log; updating the privacy description mining result based on the attention index, and performing personalized privacy description mining and information anonymization operation on the cloud service user interaction record by using the privacy description mining result after the description size updating is completed, wherein the personalized privacy description mining and information anonymization operation comprises the following steps: determining a plurality of groups of cloud service user activity logs with sequential connection, wherein the plurality of groups of cloud service user activity logs with sequential connection comprise an Xth group of cloud service user interaction records, a Yth group of cloud service user interaction records and a Yth group of cloud service user interaction records for completing individual privacy anonymization processing, X and Y are positive integers, and Y = X-1; obtaining a privacy description mining result of the X group of cloud service user interaction records based on the X group of cloud service user interaction records, the Y group of cloud service user interaction records and the Y group of cloud service user interaction records which finish individual privacy anonymization processing; and performing information anonymization operation on the X group of cloud service user interaction records by using the privacy description mining result to obtain X group of cloud service user interaction records which are subjected to individual privacy anonymization processing.

The method and the device can obtain the privacy description mining result of the X group of cloud service user interaction records in the cloud service user activity log, then utilize the privacy description mining result of the X group of cloud service user interaction records to carry out personalized privacy description mining on the X group of cloud service user interaction records, accurately and reliably carry out anonymization processing on user privacy information in the X group of cloud service user interaction records, and obtain the X group of cloud service user interaction records which finish individual privacy anonymization processing. In this way, since the information anonymization operation is performed for individual privacy, and attention indexes (such as characteristic dimensions) of a privacy description mining result of continuous cloud service user interaction records can be considered in the privacy anonymization processing process, anonymization processing of relatively important user information can be realized as accurately as possible without anonymizing a large amount of user information, thereby reducing the overhead of system resources and improving the efficiency of user information protection.

For some design ideas that can be implemented independently, the obtaining a privacy description mining result of the X-th group of cloud service user interaction records based on the X-th group of cloud service user interaction records, the Y-th group of cloud service user interaction records, and the Y-th group of cloud service user interaction records that complete individual privacy anonymization processing includes: and performing personalized privacy description mining on service session behavior events of cloud service user interaction records to be subjected to user information protection to obtain privacy description mining results, wherein the cloud service user interaction records to be subjected to user information protection are obtained by combining the Xth group of cloud service user interaction records, the Yth group of cloud service user interaction records and the Yth group of cloud service user interaction records which finish individual privacy anonymization processing in the attention plane.

According to the design, based on the service session behavior events recorded by the interaction of the cloud service users in the Y group and the individual privacy data between the service session behavior events recorded by the interaction of the cloud service users completing individual privacy anonymization processing in the Y group, the privacy description mining result of the service session behavior events is obtained, and the privacy description mining result is used for carrying out individualized privacy description mining on the corresponding service session behavior events in the interaction records of the cloud service users in the X group so as to accurately and reliably carry out anonymization processing on the service session behavior events in the interaction records of the cloud service users in the X group; the privacy description mining result is respectively determined for each service session behavior event in the X group of cloud service user interaction records, and anonymization processing can be accurately and reliably carried out on the X group of cloud service user interaction records, so that the cloud service user interaction records which are subjected to individual privacy anonymization processing can be more accurate and reliable.

For other design ideas which can be independently implemented, the personalized privacy description mining is performed on the service session behavior event of the cloud service user interaction record to be subjected to user information protection, and a privacy description mining result is obtained, wherein the privacy description mining result comprises: performing personalized privacy description mining on the cloud service user interaction record to be subjected to user information protection to determine that the service session behavior event of the Yth group of cloud service user interaction records is compared with sensitive event content change data of the service session behavior event of the Xth group of cloud service user interaction records, and obtaining related behavior event content, wherein the sensitive event content change data comprise an event heat value and an event interaction state; and carrying out individual image recognition operation on the associated behavior event content to obtain the privacy description mining result.

By the design, based on the sensitive event content change data between the service session behavior events recorded by the interaction of the Y-th group of cloud service users and the service session behavior events recorded by the interaction of the X-th group of cloud service users, the associated behavior event content of the service session behavior events is obtained, and the associated behavior event content can be used for pairing operation in the later period. And then, carrying out personalized privacy description mining on the content of the associated behavior event to determine individual privacy data between the service session behavior event recorded by the interaction of the cloud service users in the Y th group and the service session behavior event recorded by the interaction of the cloud service users in the Y th group which completes the individual privacy anonymization processing, and obtaining a privacy description mining result, so that the privacy description mining result not only contains the individual privacy data between the service session behavior event recorded by the interaction of the cloud service users in the Y th group and the service session behavior event recorded by the interaction of the cloud service users in the Y th group which completes the individual privacy anonymization processing, but also contains sensitive event content change data between the service session behavior event recorded by the interaction of the cloud service users in the Y th group and the service session behavior event recorded by the interaction of the cloud service users in the X th group, and thus being beneficial to improving the reliability of carrying out the anonymization processing on the interaction records of the cloud service users in the X group.

For some design ideas which can be implemented independently, the performing, by using the privacy description mining result, an information anonymization operation on the X-th group of cloud service user interaction records to obtain an X-th group of cloud service user interaction records for completing individual privacy anonymization processing includes: performing personalized privacy description mining on service session behavior events of key service item privacy distribution recorded by the X group of cloud service user interaction by using the privacy description mining result to obtain first key service item privacy distribution; and carrying out information anonymization operation on the privacy distribution of the first key service item to obtain the X-th group of cloud service user interaction records which are subjected to individual privacy anonymization processing.

By the design, information anonymization operation is performed on the privacy distribution of the key service items recorded by the X group of cloud service user interaction by using the privacy description mining result, so that the resource waste in the anonymization processing process can be reduced, and the efficiency of the information anonymization operation can be obviously improved.

For still other design ideas that can be implemented independently, the performing personalized privacy description mining on the service session behavior events of the key service item privacy distribution recorded by the xth group of cloud service user interactions by using the privacy description mining result to obtain a first key service item privacy distribution includes: updating the description size of the privacy description mining result, so that the number of the attention indexes of the privacy description mining result is consistent with the number of the attention indexes of the privacy distribution of the key service items recorded by the X group of cloud service user interaction; and performing personalized privacy description mining on the service session behavior events of the key service item privacy distribution recorded by the X group of cloud service user interaction by using the privacy description mining result of which the description size is updated, so as to obtain the first key service item privacy distribution.

By means of the design, the description size of the updated privacy description mining result is consistent with the description size of the key service item privacy distribution recorded by the X group of cloud service users, and therefore personalized privacy description mining can be accurately performed on the key service item privacy distribution recorded by the X group of cloud service users by means of the updated description size privacy description mining result.

For still another design idea that can be implemented independently, after the performing personalized privacy description mining on the cloud service user interaction record to be subjected to user information protection to determine that the service session behavior event of the Y-th group of cloud service user interaction records is compared with the sensitive event content change data of the service session behavior event of the X-th group of cloud service user interaction records, and obtaining the content of the associated behavior event, the method further includes: and performing personalized privacy description mining on the service session behavior events of the key service item privacy distribution recorded by the Y-th group of cloud service user interaction by using the associated behavior event content to obtain second key service item privacy distribution.

By means of the design, personalized privacy description mining is performed on the service session behavior events of the key service item privacy distribution of the Yth group of cloud service user interaction records by using the content of the associated behavior events, and the key service item privacy distribution of the Yth group of cloud service user interaction records can be matched with the X group of cloud service user interaction records.

For still some design ideas that can be implemented independently, the performing personalized privacy description mining on the service session behavior event of the key service item privacy distribution recorded by the interaction of the Y-th group of cloud service users by using the content of the associated behavior event to obtain a second key service item privacy distribution includes: updating the description size of the associated behavior event content to enable the number of the attention indexes of the associated behavior event content to be consistent with the number of the attention indexes of the privacy distribution of the key service items of the interaction records of the Yth group of cloud service users; and performing personalized privacy description mining on the service session behavior events of the key service item privacy distribution of the interaction records of the cloud service users of the Yth group, which are subjected to individual privacy anonymization processing, by using the associated behavior event content with the updated description size, so as to obtain the second key service item privacy distribution.

By means of the design, the description size of the disassociation behavior event content is kept consistent with the description size of the key service item privacy distribution of the interaction record of the Yth group of cloud service users by updating the description size of the disassociation behavior event content, and therefore personalized privacy description mining on the key service item privacy distribution of the interaction record of the Yth group of cloud service users by utilizing the updated description size of the disassociation behavior event content is accurately and reliably achieved.

For still other design ideas that can be implemented independently, the performing information anonymization on the privacy distribution of the first key service item to obtain the xth group of cloud service user interaction records for completing individual privacy anonymization processing includes: splicing the first key service item privacy distribution and the second key service item privacy distribution to obtain a third key service item privacy distribution; and carrying out information anonymization operation on the privacy distribution of the third key service item to obtain the X group of cloud service user interaction records which are subjected to individual privacy anonymization processing.

By means of the design, the first key service item privacy distribution and the second key service item privacy distribution are spliced, the quality of anonymization processing of the X group of cloud service user interaction records can be guaranteed, and then information anonymization operation is carried out on the spliced third key service item privacy distribution to obtain the X group of cloud service user interaction records which are subjected to individual privacy anonymization processing.

For some design ideas which can be implemented independently, the personalized privacy description mining is performed on the cloud service user interaction record to be subjected to user information protection to determine that the service session behavior event of the Y-th group of cloud service user interaction records is compared with the sensitive event content change data of the service session behavior event of the X-th group of cloud service user interaction records, and obtain the content of the associated behavior event includes: combining the X group of cloud service user interaction records, the Y group of cloud service user interaction records and the Y group of cloud service user interaction records which are subjected to individual privacy anonymization processing under the attention level to obtain the cloud service user interaction records to be subjected to user information protection; carrying out individual portrait recognition operation on the cloud service user interaction record to be subjected to user information protection, and obtaining privacy distribution of a fourth key service item; performing personalized privacy description mining on the privacy distribution of the fourth key service item to obtain the privacy distribution of a fifth key service item; updating the attention index quantity of the privacy distribution of the fifth key service item to a first specified numerical value by utilizing personalized privacy description mining, and obtaining the content of the associated behavior event.

According to the design, personalized privacy description mining is carried out on the cloud service user interaction records to be subjected to user information protection, the fact that the service session behavior events of the Y-th group of cloud service user interaction records are compared with the sensitive event content change data of the service session behavior events of the X-th group of cloud service user interaction records is determined, and the personalized privacy description mining is favorably and continuously utilized subsequently to update the attention index quantity of the privacy distribution of the fifth key service item to be the first designated numerical value.

For still other design ideas which can be independently implemented, the associating behavior event content performs individual image recognition operation to obtain the privacy description mining result, and the method comprises the following steps: updating the attention index quantity of the content of the associated behavior event to a second specified numerical value by utilizing personalized privacy description mining to obtain privacy distribution of a sixth key service item; splicing the fourth key service item privacy distribution and the sixth key service item privacy distribution to obtain a seventh key service item privacy distribution; and performing personalized privacy description mining on the privacy distribution of the seventh key service item to determine that the service session behavior event of the cloud service user interaction record of which the Yth group completes individual privacy anonymization processing is compared with the individual privacy data of the service session behavior event of the Yth group cloud service user interaction record, and obtaining the privacy description mining result.

By the design, the privacy description mining result is obtained by performing personalized privacy description mining on the content of the associated behavior event, so that the privacy description mining result not only covers sensitive event content change data of the service session behavior event recorded by the Y-th group of cloud service user interaction records compared with the service session behavior event recorded by the X-th group of cloud service user interaction records, but also covers individual privacy data of the service session behavior event recorded by the Y-th group of cloud service user interaction records compared with the service session behavior event recorded by the Y-th group of cloud service user interaction records, and the quality of anonymization processing on the X-th group of cloud service user interaction records by using the privacy description mining result in a later period can be improved.

For still other design ideas that can be implemented independently, the performing personalized privacy description mining on the privacy distribution of the seventh key service item to determine that the Y-th group of cloud service user interaction records that complete individual privacy anonymization processing are compared with the individual privacy data of the service session behavior event of the Y-th group of cloud service user interaction records, and obtaining the privacy description mining result includes: performing personalized privacy description mining on the privacy distribution of the seventh key service item to obtain the privacy distribution of the eighth key service item; updating the attention index quantity of the eighth key service item privacy distribution to the first designated numerical value by utilizing personalized privacy description mining, and obtaining the privacy description mining result.

By means of the design, personalized privacy description mining is carried out on the privacy distribution of the seven key service items, and the content change data of the service session behavior events recorded by the interaction of the cloud service users in the Y-th group is determined to be compared with the content change data of the sensitive events recorded by the interaction of the cloud service users in the Y-th group, which are processed in an individual privacy anonymization mode, so that the attention index quantity of the privacy distribution of the eighth key service item can be updated to be the first designated numerical value by means of the personalized privacy description mining continuously in the follow-up process.

For still other design ideas that can be implemented independently, the performing information anonymization on the privacy distribution of the third key service item to obtain the xth group of cloud service user interaction records for completing individual privacy anonymization processing includes: performing k-anonymity privacy analysis on the privacy distribution of the third key service item to obtain the privacy distribution of a ninth key service item; performing personalized privacy description mining on the privacy distribution of the ninth key service item to obtain a cloud service user interaction record after anonymous operation of the Xth group of information; weighting an event subject label of a first service session behavior event recorded by the X group of cloud service user interaction records with an event subject label of a second service session behavior event recorded by the cloud service user interaction records after the X group of information is anonymously operated to obtain the X group of cloud service user interaction records after the individual privacy anonymization processing is completed, wherein the positioning analysis data of the first service session behavior event in the X group of cloud service user interaction records is consistent with the positioning analysis data of the second service session behavior event in the X group of cloud service user interaction records after the X group of information is anonymously operated.

According to the design, firstly, anonymous privacy analysis and personalized privacy description mining are utilized to realize the information anonymity operation of privacy distribution of a third key service item, the cloud service user interaction record after anonymous operation of the X group of information is obtained, secondly, the event subject label weighting of the service session behavior event corresponding to the X group of cloud service user interaction record after anonymous operation of the X group of information and the cloud service user interaction record after anonymous operation of the X group of information is utilized to obtain the cloud service user interaction record after anonymous processing of the individual privacy of the X group, and the anonymity processing quality can be guaranteed.

A second aspect is a smart cloud server comprising a memory and a processor; the memory and the processor are coupled; the memory for storing computer program code, the computer program code comprising computer instructions; wherein the computer instructions, when executed by the processor, cause the smart cloud server to perform the method of the first aspect.

Drawings

Fig. 1 is a schematic flowchart of a user information protection method for smart cloud services according to an embodiment of the present invention.

Fig. 2 is a block diagram of a user information protection device for smart cloud services according to an embodiment of the present invention.

Detailed Description

In the following, the terms "first", "second" and "third", etc. are used for descriptive purposes only and are not to be construed as indicating or implying relative importance or to imply that the number of indicated technical features is significant. Thus, a feature defined as "first," "second," or "third," etc., may explicitly or implicitly include one or more of that feature.

Fig. 1 illustrates a flowchart of a user information protection method for a smart cloud service according to an embodiment of the present invention, where the user information protection method for the smart cloud service may be implemented by a smart cloud server, and the smart cloud server may include a memory and a processor; the memory and the processor are coupled; the memory for storing computer program code, the computer program code comprising computer instructions; when the processor executes the computer instructions, the smart cloud server is enabled to execute the technical scheme described in the following steps.

The design idea of the user information protection method for the smart cloud service provided by the embodiment of the invention is as follows: determining a privacy description mining result of a service session behavior event of a cloud service user interaction record to be subjected to user information protection according to a cloud service user activity log; updating the privacy description mining result based on the attention index, and performing personalized privacy description mining and information anonymity operation on the cloud service user interaction record by using the privacy description mining result of which the description size is updated. The information anonymization operation is performed according to individual privacy, and attention indexes (such as characteristic dimensions) of privacy description mining results of continuous cloud service user interaction records can be considered in the privacy anonymization processing process, so that anonymization processing of relatively important user information can be accurately realized as far as possible without anonymizing a large amount of user information, the expenditure of system resources is reduced, and the efficiency of user information protection is improved.

It will be appreciated that further embodiments of the overall design concept described above can be achieved as follows.

Step 301, determining that a plurality of groups of cloud service user activity logs in sequential connection exist.

In the embodiment of the invention, the multiple groups of cloud service user activity logs with sequential connection in time sequence comprise an Xth group of cloud service user interaction records, a Yth group of cloud service user interaction records and a Yth group of cloud service user interaction records for completing individual privacy anonymization processing, wherein X and Y are positive integers, and Y = X-1.

In the embodiment of the invention, a plurality of groups of cloud service user activity logs with sequential connection can be determined by using the log detection thread. The X-th group of cloud service user interaction records and the Y-th group of cloud service user interaction records are two groups of cloud service user interaction records adjacent to each other in the cloud service user activity log, further, the X-th group of cloud service user interaction records can be understood as the last group of cloud service user interaction records of the Y-th group of cloud service user interaction records, and the X-th group of cloud service user interaction records can be understood as the group of cloud service user interaction records to be subjected to information anonymization operation at present. The cloud service user interaction records of the Yth group which are subjected to the individual privacy anonymization processing can be understood as cloud service user interaction records obtained after information anonymization operation is carried out on the cloud service user interaction records of the Yth group.

For example, the plurality of groups of cloud service user activity logs having time-series connection may be understood as a plurality of groups of continuous cloud service user activity logs. Further, the cloud service user activity log may relate to the fields of e-commerce, online office, smart city, digital medical service, cloud game, and the like, and when the foregoing fields are taken as examples to implement the embodiment of the present invention, a person skilled in the art may perform adaptive adjustment on the actual content of the cloud service user activity log, which is not limited by the embodiment of the present invention.

It can be understood that anonymization processing on the cloud service user activity log in the embodiment of the present invention may be a cyclic process, and it is simply understood that the xth group of cloud service user interaction records which complete anonymization processing on the individual privacy will be used as raw material information (which may be understood as input information) of an anonymization operation process of the xth group of cloud service user interaction record information, where V is a positive integer, and V = X +1.

For example, if X is 1, the currently anonymized interaction log is the first group in the cloud service user activity log. At this time, the Yth group of cloud service user interaction records and the Yth group of cloud service user interaction records which are subjected to individual privacy anonymization processing are both the Xth group, and 3 groups of first group of cloud service user interaction records are determined.

In the embodiment of the invention, a set obtained by arranging each group of cloud service user interaction records in the cloud service user activity log according to the collected time sequence is called a cloud service user activity log set. And the cloud service user interaction record obtained after the information anonymization operation is called as the cloud service user interaction record for completing individual privacy anonymization processing.

According to the embodiment of the invention, the information anonymity operation is carried out on the cloud service user activity log according to the set in the cloud service user activity log, and the information anonymity operation is carried out only on a group of cloud service user interaction records in each round.

It can be understood that the cloud service user activity log and the cloud service user interaction record completing the individual privacy anonymization processing can be recorded in the cloud space, wherein the cloud service user activity log refers to a dynamic and streaming user activity log, that is, the cloud service user activity log is recorded in the cloud space according to the rule of the cloud service user activity log set. Therefore, the smart cloud server can directly determine the X-th group of cloud service user interaction records, the Y-th group of cloud service user interaction records and the Y-th group of cloud service user interaction records which are processed in an individual privacy anonymization mode from the cloud space.

It can be understood that, in the embodiment of the present invention, the cloud service user activity log may be a cloud service user activity log obtained through a log detection thread of the smart cloud server, or may also be a cloud service user activity log recorded in a cloud space, but is not limited thereto.

Step 302, based on the xth group of cloud service user interaction records, the yth group of cloud service user interaction records, and the yth group of cloud service user interaction records for completing individual privacy anonymization processing, obtaining a privacy description mining result of the xth group of cloud service user interaction records.

In the embodiment of the invention, the X-th group of cloud service user interaction records, the Y-th group of cloud service user interaction records and the Y-th group of cloud service user interaction records which are subjected to individual privacy anonymization processing are combined (for example, superposed) in an attention level (different attention indexes correspond to different attention dimensions), and the cloud service user interaction records to be subjected to user information protection are obtained. For example, if the scales of the X-th group of cloud service user interaction records, the Y-th group of cloud service user interaction records, and the Y-th group of cloud service user interaction records for completing the individual privacy anonymization processing are all a × b, and the scale of the cloud service user interaction records to be subjected to the user information protection obtained after the combination is a × a 3b, in other words, the number of service session behavior events in the cloud service user interaction records to be subjected to the user information protection obtained after the combination of the 3 groups of cloud service user interaction records (the X-th group of cloud service user interaction records, the Y-th group of cloud service user interaction records, and the Y-th group of cloud service user interaction records for completing the individual privacy anonymization processing) is kept unchanged relative to the number of service session behavior events in one group of cloud service user interaction records in the 3 groups of cloud service user interaction records, but the amount of attention index of each service session behavior event is adjusted to be three times that in one group of cloud service user interaction records in the 3 groups of cloud service user interaction records.

In the embodiment of the present invention, personalized privacy description mining on service session behavior events of cloud service user interaction records to be subjected to user information protection may be implemented by a plurality of combined privacy description mining submodels (convolutional layers).

The method comprises the steps of utilizing the service session behavior events of the cloud service user interaction records to be subjected to user information protection to carry out personalized privacy description mining, determining key event contents (such as characteristic information) of the service session behavior events in the cloud service user interaction records to be subjected to user information protection, and obtaining privacy description mining results. The key event content includes sensitive event content change data (for example, a front-back difference condition of event content with strong correlation with privacy information) of a service session behavior event recorded by a Y-th group of cloud service user interaction records compared with the service session behavior event recorded by the X-th group of cloud service user interaction records, and individual privacy data of the service session behavior event recorded by the Y-th group of cloud service user interaction records compared with the service session behavior event recorded by the Y-th group of cloud service user interaction records subjected to individual privacy anonymization processing. The sensitive event content change data includes a sensitive event heat value (such as attention or activity of an event) and a sensitive event interaction state (such as whether the event is in an interaction process, a pause process or a termination process) of a service session behavior event in the Y-th group of cloud service user interaction records relative to a corresponding service session behavior event in the X-th group of cloud service user interaction records.

It can be understood that the privacy description mining result in the embodiment of the present invention may be understood as a result obtained by performing personalized privacy description mining on the cloud service user interaction record to be subjected to user information protection, and the result is used as a privacy description mining unit for personalized privacy description mining in the later mining of the present invention.

The personalized privacy description mining of the service session behavior events of the cloud service user interaction record to be subjected to user information protection can be understood as the personalized privacy description mining of each service session behavior event of the cloud service user interaction record to be subjected to user information protection, and the privacy description mining result of each service session behavior event is obtained respectively. In combination with other examples, the cloud service user interaction record to be subjected to user information protection is of a × 3b scale, in other words, the cloud service user interaction record to be subjected to user information protection includes a × a service session behavior events, and after the service session behavior events of the cloud service user interaction record to be subjected to user information protection are subjected to personalized privacy description mining, a set of key service item privacy distributions of a × a can be obtained, where each service session behavior event in the key service item privacy distributions of a may be used as a privacy description mining result of performing information anonymization operation on the service session behavior events in the X-th set of cloud service user interaction records at a later stage.

And 303, performing information anonymization operation on the X-th group of cloud service user interaction records by using the privacy description mining result to obtain X-th group of cloud service user interaction records for completing individual privacy anonymization processing.

Step 3031, performing personalized privacy description mining on the service session behavior events of the key service item privacy distribution recorded by the X group of cloud service user interaction by using the privacy description mining result to obtain first key service item privacy distribution.

In the embodiment of the present invention, the key service item privacy distribution of the xth group of cloud service user interaction records may be obtained by performing a key content identification operation on the xth group of cloud service user interaction records, where the key content identification operation may be personalized privacy description mining or a downsampling operation, and the embodiment of the present invention does not limit this. Further, the key service item privacy distribution can be understood as feature distribution or feature graph of related user information privacy of the key service item, and the key service item privacy distribution can completely and comprehensively record the related user information privacy of the key service item as far as possible, so that reliable analysis basis is provided for subsequent privacy anonymity.

And obtaining a privacy description mining result of each service session behavior event in the cloud service user interaction record to be subjected to user information protection by utilizing the processing in the step 302, wherein the number of the service session behavior events in the cloud service user interaction record to be subjected to user information protection is consistent with the number of the service session behavior events in the X-th group of cloud service user interaction records, and the service session behavior events in the cloud service user interaction record to be subjected to user information protection correspond to the service session behavior events in the X-th group of cloud service user interaction records one by one. In the embodiments of the present invention, reference may be made to the following embodiments for respective further explanations: the service session behavior event _ a in the cloud service user interaction record to be subjected to user information protection corresponds to the service session behavior event _ B in the X-th group of cloud service user interaction records one by one, in other words, the positioning analysis data of the service session behavior event _ a in the cloud service user interaction record to be subjected to user information protection is consistent with the positioning analysis data of the service session behavior event _ B in the X-th group of cloud service user interaction records. By way of example, the location resolution data may be understood as the relative location or distribution of events in the log.

Step 3032, performing information anonymization operation on the privacy distribution of the first key service item, and obtaining the X group of cloud service user interaction records which are subjected to individual privacy anonymization processing.

In the embodiment of the invention, the information anonymization operation can be realized by utilizing k-anonymity privacy analysis, and can also be obtained by utilizing the k-anonymity privacy analysis and personalized privacy description mining and merging.

It can be understood that, in order to significantly guarantee the quality of anonymization processing on the X-th group of cloud service user interaction records, the event topic tag of the service session behavior event in the cloud service user interaction record obtained by performing information anonymization operation on the privacy distribution of the first key service item is weighted with the event topic tag of the service session behavior event in the X-th group of cloud service user interaction records, and the cloud service user interaction record obtained after "weighting" is used as the X-th group of cloud service user interaction records for completing individual privacy anonymization processing. And obtaining the X group of cloud service user interaction records for completing individual privacy anonymization processing by utilizing the weighting information and utilizing the X group of cloud service user interaction records.

For example, if the event topic tag of the service session behavior event _ C in the cloud service user interaction record obtained after the information anonymization operation is 100, and the event topic tag of the service session behavior event _ D in the xth group of cloud service user interaction records is 75, the event topic tag of the service session behavior event _ E in the cloud service user interaction record obtained after the "weighting" that the xth group completes the individual privacy anonymization processing is 175. Positioning analysis data of the service session behavior event _ C in a cloud service user interaction record to be subjected to user information protection, positioning analysis data of the service session behavior event _ D in an X group of cloud service user interaction records, and positioning analysis data of the service session behavior event _ E in the X group of cloud service user interaction records subjected to individual privacy anonymization processing are consistent.

Based on the above contents, the content change sets of the sensitive events of different service session behavior events in the cloud service user interaction records with differences have differences, and the more complicated the content change set of the sensitive events of the service session behavior events is, the greater the privacy protection importance of the content change set of the sensitive events is. Because the privacy protection urgency degrees of different service session behavior events in the different cloud service user interaction records are different, the corresponding privacy description mining results are matched for the different service session behavior events, so that the information anonymity protection of each service session behavior event can be better realized, further the individualized anonymity protection processing of the different cloud service user interaction records is realized, and the leakage of the user privacy information in the service interaction process is avoided.

According to the embodiment of the invention, based on the service session behavior event recorded by the interaction of the cloud service users in the Yth group and the individual privacy data between the service session behavior events recorded by the interaction of the cloud service users in the Yth group, which finish the individual privacy anonymization processing, the privacy description mining result of the service session behavior event is obtained, and the privacy description mining result is used for carrying out the individualized privacy description mining on the corresponding service session behavior event in the interaction records of the cloud service users in the Xth group so as to accurately and reliably carry out the anonymization processing on the service session behavior event in the interaction records of the cloud service users in the Xth group; the privacy description mining result is respectively determined for each service session behavior event in the X group of cloud service user interaction records, anonymization processing can be accurately and reliably carried out on the X group of cloud service user interaction records, the cloud service user interaction records which are subjected to individual privacy anonymization processing can be more accurate and reliable, and meanwhile, the efficiency in the anonymization operation process of all information can be improved.

For other design considerations that can be implemented independently, the content recorded in step 3031 can be implemented in the following manner.

Step 401, performing personalized privacy description mining on a cloud service user interaction record to be subjected to user information protection to determine that a service session behavior event of a Y-th group of cloud service user interaction records is compared with sensitive event content change data of the service session behavior event of an X-th group of cloud service user interaction records, and obtaining associated behavior event content, wherein the sensitive event content change data includes an event heat value and an event interaction state.

In the embodiment of the present invention, the sensitive event content change data includes an event heat value and an event interaction state, and the sensitive event content change data that can be understood as a service session behavior event refers to a sensitive event content change set in time information from a Y-th group of sequential nodes (obtaining time information recorded by a Y-th group of cloud service user interactions) to an X-th group of the service session behavior event (obtaining time information recorded by an X-th group of cloud service user interactions).

By the design, the content change data of the service session behavior event recorded by the Y-th group of cloud service user interaction records can be more beneficial to improving the efficiency of anonymization processing on the X-th group of cloud service user interaction records compared with the content change data of the sensitive event recorded by the X-th group of cloud service user interaction records.

In the embodiment of the invention, personalized privacy description mining on the service session behavior events of the cloud service user interaction records to be subjected to user information protection can be realized by a plurality of combined privacy description mining submodels, and the number of the privacy description mining submodels and the dimensionality of a privacy description mining unit in the privacy description mining submodel are not limited too much.

By utilizing the personalized privacy description mining of the service session behavior event of the cloud service user interaction record to be subjected to user information protection, the key event content of the service session behavior event in the cloud service user interaction record to be subjected to user information protection can be determined, and the associated behavior event content is obtained. The key event content comprises sensitive event content change data of service session behavior events recorded by the Y-th group of cloud service user interaction records compared with the service session behavior events recorded by the X-th group of cloud service user interaction records.

It can be understood that the associated behavior event content in the embodiment of the present invention may be understood as a result obtained by performing personalized privacy description mining on a cloud service user interaction record to be subjected to user information protection, and the result is used as a privacy description mining unit for personalized privacy description mining in a later operation of the embodiment of the present invention. In actual implementation, in view of the fact that the content of the associated behavior event is subjected to personalized privacy description mining by using the cloud service user interaction record to be subjected to user information protection, the service session behavior event of the Y-th group of cloud service user interaction records is determined to be obtained by comparing the sensitive event content change data of the service session behavior event of the X-th group of cloud service user interaction records, and therefore the service session behavior event of the X-th group of cloud service user interaction records can be paired subsequently by using the content of the associated behavior event. For example, the cloud service user interaction record to be subjected to user information protection may be understood as a cloud service user interaction record which needs to be processed currently.

It can be understood that the associated behavior event content obtained in the embodiment of the present invention is also obtained in real time, and by using the above operation thought, the associated behavior event content of each service session behavior event in the X-th group of cloud service user interaction records is obtained.

And 402, carrying out individual image recognition operation on the associated behavior event content to obtain the privacy description mining result.

In the embodiment of the invention, the individual portrait identification operation can be personalized privacy description mining or downsampling operation.

The individual portrait recognition operation can be understood as personalized privacy description mining, the personalized privacy description mining can be realized through a plurality of combined privacy description mining submodels, and the number of the privacy description mining submodels and the dimension of a privacy description mining unit in the privacy description mining submodel are not limited in the embodiment of the invention.

It is to be understood that the personalized privacy description mining in step 402 is different from the personalized privacy description mining in step 401. For example, if the personalized privacy description mining in step 401 is implemented by a privacy description mining submodel (the size dimension of the privacy description mining unit is 3c × 3c) whose number of attention indexes is "8", the personalized privacy description mining in step 402 is implemented by five privacy description mining submodels (the size dimension of the privacy description mining unit is 3c × 3c) whose number of attention indexes is "16", both of which (the three privacy description mining submodels and the five privacy description mining submodels) are actually personalized privacy description mining, but the actual implementation flows of both are different.

In view of the fact that the cloud service user interaction record to be subjected to user information protection can be obtained by combining the X-th group of cloud service user interaction records, the Y-th group of cloud service user interaction records and the Y-th group of cloud service user interaction records which are subjected to individual privacy anonymization processing in the attention plane, the cloud service user interaction record to be subjected to user information protection contains the X-th group of cloud service user interaction records, the Y-th group of cloud service user interaction records and the Y-th group of cloud service user interaction records which are subjected to individual privacy anonymization processing. The personalized privacy description mining in step 401 tends to determine sensitive event content change data of the service session behavior event recorded by the Y-th group of cloud service user interaction records compared with the service session behavior event recorded by the X-th group of cloud service user interaction records, in other words, through the processing in step 401, individual privacy data between the Y-th group of cloud service user interaction records in the cloud service user interaction records to be subjected to user information protection and the Y-th group of cloud service user interaction records subjected to individual privacy anonymization processing are not determined.

It can be understood that before the individual portrait identification operation is performed on the associated behavior event content, the cloud service user interaction record to be subjected to user information protection and the associated behavior event content can be spliced, so that the associated behavior event content obtained after splicing contains individual privacy data between the Yth group of cloud service user interaction records and the Yth group of cloud service user interaction records which are subjected to individual privacy anonymization processing.

And performing personalized privacy description mining on the content of the associated behavior event, determining that the Y-th group of cloud service user interaction records which are subjected to individual privacy anonymization processing are compared with the individual privacy data of the service session behavior event of the Y-th group of cloud service user interaction records, and obtaining a privacy description mining result. The individual privacy data may be understood as a corresponding list (for example, may be understood as a mapping relationship) between the service session behavior event recorded by the interaction of the cloud service user in the Y-th group and the service session behavior event recorded by the interaction of the cloud service user in the Y-th group that completes the individual privacy anonymization processing, in other words, a corresponding list between the service session behavior event before the information anonymization operation and the service session behavior event that completes the individual privacy anonymization processing.

Therefore, the privacy description mining result obtained by carrying out personalized privacy description mining on the content of the associated behavior event not only covers individual privacy data between the service session behavior event recorded by the interaction of the cloud service users in the Yth group and the service session behavior event recorded by the interaction of the cloud service users in the Yth group, which completes individual privacy anonymization processing, but also covers sensitive event content change data between the service session behavior event recorded by the interaction of the cloud service users in the Yth group and the service session behavior event recorded by the interaction of the cloud service users in the Xth group. And subsequently, personalized privacy description mining is carried out on the service session behavior events recorded by the X group of cloud service user interaction by using the privacy description mining result, so that the efficiency of anonymization processing can be improved.

The embodiment of the invention obtains the content of the associated behavior event of the service session behavior event based on the sensitive event content change data between the service session behavior event recorded by the interaction of the Y-th group of cloud service users and the service session behavior event recorded by the interaction of the X-th group of cloud service users, and can subsequently perform pairing operation by using the content of the associated behavior event. And then, carrying out personalized privacy description mining on the content of the associated behavior event to determine individual privacy data between the service session behavior event recorded by the interaction of the cloud service users in the Y th group and the service session behavior event recorded by the interaction of the cloud service users in the Y th group which completes individual privacy anonymization processing, and obtaining a privacy description mining result, so that the privacy description mining result not only contains the individual privacy data between the service session behavior event recorded by the interaction of the cloud service users in the Y th group and the service session behavior event recorded by the interaction of the cloud service users in the Y th group which completes individual privacy anonymization processing, but also contains sensitive event content change data between the service session behavior event recorded by the interaction of the cloud service users in the Y th group and the service session behavior event recorded by the interaction of the cloud service users in the X th group, and is favorable for improving the reliability of anonymization processing on the interaction records of the cloud service users in the X group.

According to the related technical scheme, the privacy description mining result and the content of the associated behavior event are obtained by performing personalized privacy description mining on the interaction record of the cloud service user. Because the number of service session behavior events contained in the cloud service user interaction record is large, if the cloud service user interaction record is directly processed, a large amount of resources are wasted, and the efficiency of anonymization processing is reduced at the same time, in view of this, a design idea for obtaining a privacy description mining result and associated behavior event content according to the privacy distribution of the key service item is shown below.

For other independently implementable design ideas, obtaining privacy description mining results and associated behavioral event content can be achieved in the following manner.

Step 501, combining the X-th group of cloud service user interaction records, the Y-th group of cloud service user interaction records and the Y-th group of cloud service user interaction records which are subjected to individual privacy anonymization processing in an attention plane to obtain cloud service user interaction records to be subjected to user information protection.

For a related exemplary design idea, please refer to step 302 for implementation of obtaining a cloud service user interaction record to be protected by user information, which is not described herein too much.

Step 502, performing individual portrait recognition operation on the cloud service user interaction record to be subjected to user information protection, and obtaining privacy distribution of a fourth key service item.

The individual portrait identification operation may be implemented using several concepts, for example: a moving average operation, a down-sampling operation, etc., which are not specifically limited in this embodiment of the present invention.

In some possible implementations, a related exemplary neural network model may be used to perform individual image recognition operations on a cloud service user interaction record to be subjected to user information protection, and the neural network model sequentially includes a privacy description mining submodel with an attention index number of "8" (the privacy description mining unit size dimension is 3c. 3c), two ResNet units with an attention index number of "8" (each ResNet unit includes two privacy description mining submodels, the privacy description mining unit size dimension of the privacy description mining submodel is 3c. The privacy description mining submodel with an attention index number of "16" (the privacy description mining unit size dimension is 3c. The privacy description mining submodel with an attention index number of "16" (each ResNet unit includes two privacy description mining submodels), the privacy description mining unit description mining submodel with an attention index number of "32" (each ResNet unit includes two privacy description mining submodels, the privacy description mining unit size of the description mining unit size is 3c. The privacy description mining submodel with an attention index number of "32" and the privacy description mining submodel (each ResNet unit size of the privacy description mining unit size is 3c. The privacy description mining submodel is 32).

And sequentially carrying out personalized privacy description mining on the cloud service user interaction records to be subjected to user information protection by using the neural network model to complete mining processing of the cloud service user interaction records to be subjected to user information protection, and obtaining privacy distribution of a fourth key service item. The method comprises the steps of sequentially carrying out personalized privacy description mining on cloud service user interaction records to be subjected to user information protection by utilizing a multi-layer privacy description mining submodel, determining corresponding mapping description, and finally obtaining fourth key service item privacy distribution meeting a certain scale, so that simplified processing of the cloud service user interaction records can be realized on the premise of obtaining the fourth key service item privacy distribution of the cloud service user interaction records to be subjected to user information protection, and the resource utilization rate is improved. For example, assuming that the scale of the cloud service user interaction record to be subjected to user information protection is a × b, the scale of the fourth key service item privacy distribution obtained by the neural network model performing the individual portrait recognition operation is d × d 32, where a =4d.

For some design ideas which can be implemented independently, the implementation process of the personalized privacy description mining is as follows: the privacy description mining submodel carries out personalized privacy description mining on cloud service user interaction records to be subjected to user information protection, traverses the cloud service user interaction records to be subjected to user information protection through a privacy description mining unit, multiplies service session behavior events on the cloud service user interaction records to be subjected to user information protection by description values on a corresponding privacy description mining unit, weights all multiplication results to serve as event subject labels on cloud service user interaction records corresponding to core service session behavior events of the privacy description mining unit, finally traverses all service session behavior events in the cloud service user interaction records to be subjected to user information protection, and obtains privacy distribution of a fourth key service item. It can be understood that the traversal interval of the privacy description mining submodel can be adjusted according to actual conditions.

Step 503, performing personalized privacy description mining on the fourth key service item privacy distribution to obtain fifth key service item privacy distribution.

In the embodiment of the invention, a fourth key service item privacy distribution is input to other exemplary neural network models, and the fourth key service item privacy distribution realizes the personalized privacy description mining of the fourth key service item privacy distribution through the processing of 1 privacy description mining submodel (the size dimension of a privacy description mining unit is 3c × 3c) with the attention index quantity of 32 and two ResNet units (each ResNet unit comprises two privacy description mining submodels, and the size dimension of the privacy description mining unit of the privacy description mining submodel is 3c × 3c) one by one, so that sensitive event content change data between service session behavior events recorded by interaction of a Y group of cloud service users and service session behavior events recorded by interaction of an X group of cloud service users in the fourth key service item privacy distribution is determined, and the fifth key service item privacy distribution is obtained.

It can be understood that by processing the fourth key service item privacy distribution, the scale of the interaction record of the cloud service user is not changed, and the obtained scale of the fifth key service item privacy distribution is consistent with the scale of the fourth key service item privacy distribution.

For example, the fourth key service item privacy distribution has a size d × 32, and the fifth key service item privacy distribution obtained by the processing of step 303 also has a size d × 32.

And 504, updating the attention index quantity of the privacy distribution of the fifth key service item to a first specified numerical value by utilizing personalized privacy description mining, and obtaining the content of the associated behavior event.

In the embodiment of the present invention, in order to further determine sensitive event content change data between a service session behavior event recorded by a Y-th group of cloud service user interactions and a service session behavior event recorded by an X-th group of cloud service user interactions in a fifth key service item privacy distribution, a network layer S in a related example performs personalized privacy description mining on the fifth key service item privacy distribution, and the obtained scale of the related behavior event content is d × t × e (it can be understood that the personalized privacy description mining of the network layer S is used to update the number of attention indexes of the fifth key service item privacy distribution), where t is the number of attention indexes of the fifth key service item privacy distribution, e is a positive integer, and a value of e may be 5.

For example, the size of the privacy distribution of the fifth key service item is d × 32, and the size of the associated behavior event content obtained by the processing of the above steps is d × 32 × e, that is, d × 32e2. The associated behavioral event content comprises d × 2 five members, each member comprises 32 service session behavioral events, and the positioning resolution data of different members in the first associated behavioral event content is different. Each member was of size c × 32 × e2.

Considering that the network layer S is a privacy description mining submodel, the larger the privacy description mining unit of the privacy description mining submodel is, the greater the processing pressure is. The network layer S (privacy description mining unit) is a privacy description mining submodel with the attention index number of "32" and the privacy description mining unit size dimension of c × c. And updating the number of attention indexes of the privacy distribution of the fifth key service item by using a privacy description mining submodel with the privacy description mining unit size dimension of c x c, so that the data processing pressure can be reduced as much as possible.

And 505, updating the attention index quantity of the associated behavior event content to a second specified value by utilizing personalized privacy description mining, and obtaining privacy distribution of a sixth key service item.

Whereas the number of attention indicators of the fifth key service item privacy distribution is updated by the personalized privacy description mining in step 504, the number of attention indicators of the associated behavioral event content needs to be updated to a second specified value (which can be understood as the number of attention indicators of the fifth key service item privacy distribution) before the personalized privacy description mining is performed on the associated behavioral event content to obtain the privacy description mining result.

It can be understood that the personalized privacy description mining is used for updating the attention index quantity of the associated behavior event content to a second specified value, and a sixth key service item privacy distribution is obtained. Illustratively, the personalized privacy description mining can be implemented by using a privacy description mining submodel with the attention index number of 32 and the privacy description mining unit size dimension of c × c.

Step 506, the fourth key service item privacy distribution and the sixth key service item privacy distribution are combined under the attention level, and a seventh key service item privacy distribution is obtained.

Steps 502-504 are more intended to determine sensitive event content change data between the service session behavior event recorded by the Y-th group of cloud service user interaction records in the cloud service user interaction record to be user information protected and the service session behavior event recorded by the X-th group of cloud service user interaction records. In view of the fact that post-processing needs to determine individual privacy data between service session behavior events recorded by a Yth group of cloud service user interaction records in cloud service user interaction records to be subjected to user information protection and service session behavior events recorded by a Yth group of cloud service user interaction records to complete individual privacy anonymization processing, before post-processing, the fourth key service item privacy distribution and the sixth key service item privacy distribution are spliced to add the individual privacy data between the service session behavior events recorded by the Yth group of cloud service user interaction records and the service session behavior events recorded by the Yth group of cloud service user interaction records to complete individual privacy anonymization processing in the key service item privacy distribution.

In some examples, a splicing operation (split processing) is performed on the fourth key service item privacy distribution and the sixth key service item privacy distribution, in other words, the fourth key service item privacy distribution and the sixth key service item privacy distribution are combined under the attention level, so as to obtain a seventh key service item privacy distribution.

And 507, performing personalized privacy description mining on the privacy distribution of the seventh key service item to determine that the service session behavior event recorded by the interaction of the cloud service users, which completes individual privacy anonymization processing, of the Yth group is compared with the individual privacy data of the service session behavior event recorded by the interaction of the cloud service users of the Yth group, and obtaining the privacy description mining result.

It can be understood that the seventh key service item privacy distribution includes individual privacy data between the service session behavior event recorded by the interaction of the Y-th group of cloud service users and the service session behavior event recorded by the interaction of the Y-th group of cloud service users, which completes individual privacy anonymization processing, and the individual privacy data between the service session behavior event recorded by the interaction of the Y-th group of cloud service users and the service session behavior event recorded by the interaction of the Y-th group of cloud service users, which completes individual privacy anonymization processing, can be further determined by performing personalized privacy description mining on the seventh key service item privacy distribution, so as to obtain a privacy description mining result.

For one possible embodiment, for the above privacy description mining results obtained, examples may include the following: performing personalized privacy description mining on the privacy distribution of the seventh key service item to obtain the privacy distribution of the eighth key service item; and updating the attention index quantity of the privacy distribution of the eighth key service item to a first specified numerical value by utilizing personalized privacy description mining, and obtaining a privacy description mining result.

In some examples, a seventh key service item privacy distribution is input to a corresponding neural network model, the seventh key service item privacy distribution is processed through 1 privacy description mining submodel (the privacy description mining submodel has a size dimension of 3c × 3c) with an attention index number of 32 and two ResNet units (each ResNet unit comprises two privacy description mining submodels, and the privacy description mining unit of the privacy description mining submodel has a size dimension of 3c × 3c) one by one to achieve personalized privacy description mining of the seventh key service item privacy distribution, and determine individual privacy data between service session behavior events recorded by a Y group of cloud service user interaction in the seventh key service item privacy distribution and service session behavior events recorded by a Y group of cloud service user interaction completing individual privacy anonymization processing, so as to obtain an eighth key service item distribution.

It is understood that the architecture of the related neural network model can be similar, but those skilled in the art can flexibly adjust the model parameters according to actual requirements.

It is to be understood that the privacy description mining result obtained in step 507 is a privacy description mining result including each service session behavior event in the seventh key service item privacy distribution, and the scale of the privacy description mining unit of each service session behavior event is c × te2.

For example, the scale of the seventh key service item privacy distribution is d × 32 × e, in other words, the seventh key service item privacy distribution includes d × 2 five service session behavior events, and accordingly, the obtained privacy description mining result (with the size of d × 32e 2) includes d × 2 five privacy description mining results (i.e., each service session behavior event corresponds to one privacy description mining result, and the scale of the privacy description mining result of each service session behavior event is c × c 32e 2).

And combining the three pieces of information of the description size of each service session behavior event in the seventh key service item privacy distribution into one piece of information of the description size, and combining the information of each service session behavior event in the seventh key service item privacy distribution into one privacy description mining unit, namely a privacy description mining result of each service session behavior event.

According to the method and the device, personalized privacy description mining is carried out on the key service item privacy distribution of the cloud service user interaction record to be subjected to user information protection, sensitive event content change data between the service session behavior event recorded by the Y-th group of cloud service user interaction records and the service session behavior event recorded by the X-th group of cloud service user interaction records are determined, and the associated behavior event content of each service session behavior event is obtained. And performing personalized privacy description mining on the privacy distribution of the seventh key service item, determining individual privacy data between the service session behavior events recorded by the interaction of the cloud service users in the Yth group and the service session behavior events recorded by the interaction of the cloud service users in the Yth group which completes individual privacy anonymization processing, and obtaining a privacy description mining result of each service session behavior event. And carrying out information anonymity operation on the X-th group of cloud service user interaction records by using the associated behavior event content and the privacy description mining result.

It can be understood that, the above design solutions respectively explain how to obtain the privacy description mining result and the associated behavioral event content, and how to perform anonymization processing on the xth group of cloud service user interaction records by using the privacy description mining result and the associated behavioral event content, and obtain the xth group of cloud service user interaction records for completing individual privacy anonymization processing.

For an independently implementable design idea, the user information protection method for the smart cloud service can be further implemented through the following steps.

Step 901, performing personalized privacy description mining on service session behavior events of key service item privacy distribution recorded by the X group of cloud service user interaction by using a privacy description mining result to obtain first key service item privacy distribution.

In the embodiment of the present invention, the key service item privacy distribution of the xth group of cloud service user interaction records may be obtained by performing a key content identification operation on the xth group of cloud service user interaction records, where the key content identification operation may be personalized privacy description mining or downsampling operation, and the embodiment of the present invention does not limit this.

For some design ideas which can be independently implemented, the neural network model can be used for performing key content identification operation on the X group of cloud service user interaction records to obtain key service item privacy distribution of the X group of cloud service user interaction records. Related design considerations may be incorporated into step 502 and will not be described in great detail herein.

The neural network model is utilized to perform key content identification operation on the X group of cloud service user interaction records, the scale of key service item privacy distribution of the obtained X group of cloud service user interaction records is smaller than that of the X group of cloud service user interaction records, and the key service item privacy distribution of the X group of cloud service user interaction records contains information of the X group of cloud service user interaction records, so that the processing timeliness can be improved by subsequently processing the key service item privacy distribution of the X group of cloud service user interaction records.

Based on the description content, performing personalized privacy description mining on each service session behavior event in the cloud service user interaction record to be subjected to user information protection, respectively obtaining privacy description mining results of each service session behavior event, performing personalized privacy description mining on the service session behavior events in the key service item privacy distribution of the Xth group of cloud service user interaction records by using the privacy description mining results, wherein the privacy description mining results of each service session behavior event in the obtained privacy description mining results can be understood as being used as a privacy description mining unit of the corresponding service session behavior event in the key service item privacy distribution of the Xth group of cloud service user interaction records, and performing personalized privacy description mining on each service session behavior event in the key service item privacy distribution of the Xth group of cloud service user interaction records.

Referring to the description in step 507 above, the privacy description mining result of each service session behavior event in the privacy description mining result includes information of each service session behavior event in the seventh key service item privacy distribution, and the information is the one-attention index information in the privacy description mining result. And the service session behavior events of the key service item privacy distribution recorded by the interaction of the Xth group of cloud service users are the three-attention index information, so that the description size of the privacy description mining result needs to be updated for performing personalized privacy description mining by using the information of each service session behavior event in the seventh key service item privacy distribution as the privacy description mining unit of each service session behavior event in the key service item privacy distribution recorded by the interaction of the Xth group of cloud service users.

Taking the above into consideration, the content recorded in step 901 can be realized by the following steps: updating the description size of the privacy description mining result to enable the number of the attention indexes of the privacy description mining result to be consistent with the number of the attention indexes of the privacy distribution of the key service items recorded by the X group of cloud service user interaction; and carrying out personalized privacy description mining on the service session behavior events of the key service item privacy distribution recorded by the X group of cloud service user interaction by using the privacy description mining result of which the description size is updated, so as to obtain the first key service item privacy distribution.

Illustratively, the privacy description mining result of each service session behavior event in the obtained privacy description mining result may be used as a privacy description mining unit of the service session behavior event corresponding to the key service item privacy distribution recorded by the xth group of cloud service user interactions, and the personalized privacy description mining may be performed on the service session behavior event.

Further, the updating description size may be understood as a description size of the privacy description mining result of each service session behavior event in the privacy description mining result, and the description size of the privacy description mining result of each service session behavior event is updated from c × te2 to t × e.

For example, the scale of the privacy description mining result of each service session behavior event is c × 32e2, and after the description size of the privacy description mining result of each service session behavior event is updated, the scale of the obtained privacy description mining unit is 32 × e.

And updating and obtaining a privacy description mining result of each service session behavior event of the key service item privacy distribution recorded by the X-th group of cloud service users by using the description size, and respectively carrying out personalized privacy description mining on each service session behavior event by using the privacy description mining result of each service session behavior event, so that the anonymization processing of each service session behavior event of the key service item privacy distribution recorded by the X-th group of cloud service users can be realized, and finally the first key service item privacy distribution is obtained.

And 902, performing personalized privacy description mining on the service session behavior event of the key service item privacy distribution recorded by the Yth group of cloud service user interaction which completes individual privacy anonymization processing by using the associated behavior event content to obtain second key service item privacy distribution.

And step 901, the obtained privacy description mining result is used as the privacy description mining result of each service session behavior event of the key service item privacy distribution of the X-th group of cloud service user interaction records, the key service item privacy distribution of the X-th group of cloud service user interaction records is subjected to the same information anonymization operation, the description size of the associated behavior event content of each service session behavior event in the obtained associated behavior event content is updated to 32 × e by using the description size update in the relevant neural network model, and the service session behavior events corresponding to the key service item privacy distribution of the Y-th group of cloud service user interaction records, the individual privacy anonymization processing of which is completed, are subjected to personalized privacy description mining by using the associated behavior event content of which the description size update is completed. The key service item privacy distribution of the cloud service user interaction records of which the Yth group completes the individual privacy anonymization processing is paired with reference to the current group, namely, the positioning analysis data of each service session behavior event in the key service item privacy distribution of the cloud service user interaction records of which the Yth group completes the individual privacy anonymization processing is respectively updated according to sensitive event content change data contained in the associated behavior event content of each service session behavior event, and the second key service item privacy distribution is obtained.

The key service item privacy distribution of the cloud service user interaction records of which the Yth group completes the individual privacy anonymization processing contains a service session behavior event with certain privacy protection requirements, but the service session behavior event in the key service item privacy distribution of the cloud service user interaction records of which the Yth group completes the individual privacy anonymization processing is different from the service session behavior event of the current group. In view of this, the operation process in step 902 is used to update the positioning analysis data of the service session behavior event of the key service item privacy distribution recorded by the Y-th group of cloud service user interaction that completes the individual privacy anonymization processing, so that the service session behavior event after updating the positioning analysis data is more likely to the positioning analysis data of the X-th group of sequential nodes. In this way, later-stage operation can perform anonymization processing on information in the interaction records of the Xth group of cloud service users according to the information of the privacy distribution of the second key service item.

And 903, splicing the first key service item privacy distribution and the second key service item privacy distribution to obtain third key service item privacy distribution.

In the embodiment of the invention, by splicing the first key service item privacy distribution and the second key service item privacy distribution, the anonymization processing quality can be improved by the information of the key service item privacy distribution recorded by the interaction of the Y group cloud service users (after pairing is completed) on the premise of carrying out anonymization operation on the individual privacy data information between the sensitive event content change data between the service session behavior event recorded by the interaction of the Y group cloud service users and the service session behavior event recorded by the interaction of the X group cloud service users and the individual privacy data information between the service session behavior event recorded by the interaction of the Y group cloud service users and the service session behavior event recorded by the interaction of the Y group cloud service users completing individual privacy anonymization processing.

In some examples, the first key service item privacy profile and the second key service item privacy profile are combined under attention level (split), resulting in a third key service item privacy profile.

And 904, performing information anonymization operation on the privacy distribution of the third key service item to obtain the X group of cloud service user interaction records for completing individual privacy anonymization processing.

In this embodiment of the present invention, the information anonymization operation may be one of k-anonymity privacy analysis, reverse personalized privacy description mining, upsampling expansion, and the like, or may be a combination of the above algorithms, which is not limited in this embodiment of the present invention.

For some design ideas which can be implemented independently, a privacy anonymity submodel sequentially covers a privacy description replacement submodel with the attention index number of 16 (the size dimension of the privacy description mining unit is 3c × 3 c), two ResNet units with the attention index number of 16 (each ResNet unit comprises two privacy description mining submodels, the size dimension of the privacy description mining unit of the privacy description mining submodel is 3c × 3 c), a privacy description replacement submodel with the attention index number of 8 (the size dimension of the privacy description mining unit is 3c × 3 c), and two ResNet units with the attention index number of 8 (each ResNet unit comprises two privacy description mining submodels, and the size of the privacy description mining unit of the privacy description mining submodel is 3c × 3 c). The information anonymization operation is performed on the privacy distribution of the third key service item by using the privacy anonymization submodel, and the example of obtaining the cloud service user interaction record of the xth group which completes the individual privacy anonymization processing may include the following steps: performing k-anonymity privacy analysis on the privacy distribution of the third key service item to obtain the privacy distribution of the ninth key service item; and performing personalized privacy description mining on the privacy distribution of the ninth key service item to obtain the interaction record of the cloud service user after the anonymous operation of the Xth group of information.

It can be understood that, after the cloud service user interaction record after the X-th group information anonymization operation is obtained, the event topic tag of the first service session behavior event of the X-th group cloud service user interaction record and the event topic tag of the second service session behavior event of the cloud service user interaction record after the X-th group information anonymization operation may also be weighted, so as to obtain the cloud service user interaction record after the X-th group completes the individual privacy anonymization processing, where the positioning resolution data of the first service session behavior event in the X-th group cloud service user interaction record is consistent with the positioning resolution data of the second service session behavior event in the cloud service user interaction record after the X-th group information anonymization operation. The cloud service user interaction records of the X group completing individual privacy anonymization processing can be guaranteed to be more comprehensive.

According to the embodiment of the invention, the privacy description mining result obtained by the method can be used for carrying out information anonymity operation on the privacy distribution of the key service items recorded by the interaction of the Xth group of cloud service users, and the obtained associated behavior event content is used for carrying out pairing operation on the privacy distribution of the key service items recorded by the interaction of the Yth group of cloud service users. The method has the advantages that the information anonymization operation is carried out by utilizing the third key service item privacy distribution obtained after the first key service item privacy distribution obtained by the information anonymization operation and the second key service item privacy distribution obtained by the pairing operation are spliced, so that the quality of anonymization processing of the X group of cloud service user interaction records can be improved, and the X group of cloud service user interaction records which are subjected to individual privacy anonymization processing can be more comprehensive. In addition, the processing targets of the information anonymization operation and the pairing operation are key service item privacy distribution, so that resource waste can be avoided, and the quality of anonymization processing on the cloud service user activity log can be guaranteed.

The invention also provides an anonymization processing model of the cloud service user activity log, which is used for realizing the method. Exemplary anonymization processing models for cloud service user activity logs include: the system comprises an information characteristic identification unit, a privacy description mining unit, an associated behavior event content determining unit and a privacy anonymity submodel.

It can be understood that the relevant model network may perform adaptive adjustment and update, and the relevant model unit may perform personalized privacy description mining (i.e., pairing operation and information anonymization operation) on the service session behavior event of the key service item privacy distribution recorded by the Y-th group of cloud service user interaction and the service session behavior event of the key service item privacy distribution recorded by the X-th group of cloud service user interaction by using the privacy description dynamic mining model, to obtain the key service item privacy distribution after matching the key service item privacy distribution recorded by the Y-th group of cloud service user interaction and the key service item privacy distribution recorded by the X-th group of cloud service user interaction, to complete the key service item privacy processing of the individual privacy anonymization.

And counting the key service item privacy distribution after the pairing is completed and the key service item privacy distribution after the individual privacy anonymization processing is completed by using SPLICE under the attention level, obtaining the key service item privacy distribution after the X group completes the splicing, loading the key service item privacy distribution after the X group completes the splicing to a privacy anonymization submodel, and using the privacy anonymization processing model as a cloud service user activity log as a raw material for processing the V group cloud service user interaction records.

And using a privacy anonymity submodel to perform information anonymity operation on spliced key service items of the Xth group, obtaining cloud service user interaction records after the information anonymity operation of the Xth group, weighting event subject labels of first service session behavior events recorded by the X group of cloud service user interaction records with event subject labels of second service session behavior events recorded by the X group of cloud service user interaction records after the information anonymity operation of the X group, and obtaining cloud service user interaction records after the individual privacy anonymization processing of the X group, wherein positioning analysis data of the first service session behavior events in the X group of cloud service user interaction records are consistent with positioning analysis data of the second service session behavior events in the X group of cloud service user interaction records after the information anonymity operation of the X group. And taking the X group of cloud service user interaction records and the X group of cloud service user interaction records which are subjected to individual privacy anonymization as raw materials for processing the V group of cloud service user interaction records by using an anonymization processing model of the cloud service user activity log.

Based on the above contents, it can be known that the anonymization processing model of the cloud service user activity log needs multiple types of raw materials for performing information anonymization operation on each group of cloud service user interaction records in the cloud service user activity log, taking the X-th group of cloud service user interaction records as an example, the multiple types of raw materials are respectively: the key service item privacy distribution of the Y group of cloud service user interaction records, the Y group of cloud service user interaction records for completing individual privacy anonymization processing, the X group of cloud service user interaction records and the Y group of cloud service user interaction records for completing individual privacy anonymization processing (namely the X group of key service item privacy distribution after completing splicing).

The anonymization processing model of the cloud service user activity log provided by the embodiment of the invention can be used for carrying out information anonymization operation on the cloud service user activity log, and the cloud service user interaction record for completing individual privacy anonymization processing can be obtained only by various raw materials in the whole processing process, so that the information anonymization efficiency can be effectively improved. The privacy description mining result and the associated behavior event content are generated for each service session behavior event in the cloud service user interaction log by the privacy description mining result determining unit and the associated behavior event content determining unit, so that the anonymization quality of the anonymization processing model of the cloud service user activity log on the information anonymization quality of the cloud service user interaction log with difference in the cloud service user activity log can be improved.

Further, the training method for the relevant neural network model may be implemented according to a set cost (such as a mean square error loss function), for example, determining a difference between a cloud service user interaction record in which an X-th group output by an anonymization processing model of a cloud service user activity log completes individual privacy anonymization processing and an actual cloud service user interaction record in the X-th group output by the cloud service user interaction model (i.e., an authentication example of the X-th group output by the cloud service user interaction record), so as to construct a cost algorithm, and then performing reverse adjustment based on the cost algorithm to implement training of the neural network model.

Under some independently implementable design considerations, after obtaining the X group of cloud service user interaction records that complete the individual privacy anonymization processing, the method may further include: collecting session access logs of cloud service user interaction records which are subjected to individual privacy anonymization processing aiming at the Xth group, and identifying the session access logs to obtain risk behavior characteristics; and generating and issuing interactive warning information based on the risk behavior characteristics.

For example, abnormal intention tendency (such as illegal privacy access behavior characteristics) can be identified through the collected session access log to obtain risk behavior characteristics, so that interaction warning information can be generated by combining the risk behavior characteristics and issued to the user side to prompt the user side to actively protect individual privacy information in the interaction process, and the user side is prevented from being inducibly disclosed with relevant individual privacy information in the interaction process.

Under some design ideas which can be independently implemented, identifying the session access log to obtain risk behavior characteristics can be realized through the following technical scheme: determining, by using the collected key behavior node attributes with indication functions, a first key behavior node attribute distribution with indication functions, where a quantitative adaptation index between the collected key behavior node attributes with indication functions and behavior event expressions of the session access log reaches a first target quantitative adaptation index constraint, and a second key behavior node attribute distribution with indication functions, where the quantitative adaptation index between the collected key behavior node attributes with indication functions and behavior event expressions of the associated session access log reaches the first target quantitative adaptation index constraint; the first key behavior node attribute with the indicating function is distributed based on at least one first key behavior node attribute with the indicating function in each key behavior node attribute with the indicating function; the behavior event expression of the associated session access log is the behavior event expression of the associated session access log in the collected behavior event expressions of the associated session access logs, and the second key behavior node attribute with the indication function is distributed based on at least one second key behavior node attribute with the indication function in the key behavior node attributes with the indication function; respectively determining distribution quantitative adaptation indexes between the first key behavior node attribute distribution with the indication function and each second key behavior node attribute distribution with the indication function, and obtaining event expression quantitative adaptation indexes between the behavior event expression of the session access logs and the behavior event expression of each associated session access log; and determining the risk behavior characteristics of the session access log according to the event expression quantitative adaptation indexes.

For example, the quantitative adaptation indexes can be understood as similarity, the key behavior node attribute with the indicating function is a basic local behavior feature, and the behavior event is expressed as a global behavior feature.

Under some independently implementable design ideas, before determining, from the collected key behavior node attributes with the indication function, a first key behavior node attribute distribution with the indication function that a quantitative adaptation index between the collected key behavior node attributes and a behavior event expression of the session access log reaches a first target quantitative adaptation index constraint, the method further includes: respectively converting the behavior event expression of each associated session access log into a plurality of associated key behavior node attributes with the same statistical result, and obtaining associated key behavior node attribute distribution respectively corresponding to the behavior event expression of each associated session access log; wherein each associated key behavior node attribute in the associated key behavior node attribute distribution is adjusted according to a detection tag of each associated key behavior node attribute in a behavior event expression of a corresponding associated session access log; and determining at least one corresponding key behavior node attribute with an indication function based on the associated key behavior node attributes on the same detection label in the distribution of the associated key behavior node attributes, and obtaining the collected key behavior node attributes with the indication function. By the design, the integrity and the credibility of the collected key behavior node attributes with the indicating function can be guaranteed.

Based on the same inventive concept, fig. 2 shows a block diagram of a user information protection apparatus for smart cloud services according to an embodiment of the present invention, and the user information protection apparatus for smart cloud services may include the following modules for implementing the relevant method steps shown in fig. 1.

The log determining module 21 is configured to determine multiple groups of cloud service user activity logs having sequential connection.

The privacy mining module 22 is configured to obtain a privacy description mining result of the X-th group of cloud service user interaction records based on the X-th group of cloud service user interaction records, the Y-th group of cloud service user interaction records, and the Y-th group of cloud service user interaction records that are processed for individual privacy anonymization.

And the information processing module 23 is configured to perform information anonymization on the X-th group of cloud service user interaction records by using the privacy description mining result, and obtain an X-th group of cloud service user interaction records on which individual privacy anonymization processing is completed.

The related embodiment applied to the invention can achieve the following technical effects: the privacy description mining result of the X group of cloud service user interaction records in the cloud service user activity log can be obtained, then the privacy description mining result of the X group of cloud service user interaction records is utilized to conduct personalized privacy description mining on the X group of cloud service user interaction records, the user privacy information in the X group of cloud service user interaction records can be accurately and reliably conducted anonymization processing, and the X group of cloud service user interaction records which are used for completing individual privacy anonymization processing are obtained. In this way, since the information anonymization operation is performed for individual privacy, and attention indexes (such as characteristic dimensions) of the privacy description mining result of continuous cloud service user interaction records can be considered in the privacy anonymization processing process, anonymization processing of relatively important user information can be realized as accurately as possible without anonymizing a large amount of user information, thereby reducing the overhead of system resources and improving the efficiency of user information protection.

The above description is only a specific embodiment of the present invention. The technical field of the present invention provides a method for controlling a mobile terminal, which is capable of providing a mobile terminal with a plurality of functions, and the mobile terminal is capable of providing a mobile terminal with a plurality of functions.

Claims

1. A user information protection method for smart cloud service is applied to a smart cloud server, and the method at least comprises the following steps:

determining a privacy description mining result of a service session behavior event of a cloud service user interaction record to be subjected to user information protection according to a cloud service user activity log;

updating the privacy description mining result based on the attention index, and performing personalized privacy description mining and information anonymity operation on the cloud service user interaction record by using the privacy description mining result of which the description size is updated;

the method for determining the privacy description mining result of the service session behavior event of the cloud service user interaction record to be subjected to user information protection according to the cloud service user activity log comprises the following steps:

determining a plurality of groups of cloud service user activity logs with sequential connection, wherein the plurality of groups of cloud service user activity logs with sequential connection comprise an Xth group of cloud service user interaction records, a Yth group of cloud service user interaction records and a Yth group of cloud service user interaction records for completing individual privacy anonymization, X and Y are positive integers, and Y = X-1;

and carrying out personalized privacy description mining on the service session behavior event of the cloud service user interaction record to be subjected to user information protection to obtain a privacy description mining result, wherein the cloud service user interaction record to be subjected to user information protection is obtained by combining the X group of cloud service user interaction records, the Y group of cloud service user interaction records and the Y group of cloud service user interaction records which are subjected to individual privacy anonymization processing in the attention level.

2. The method of claim 1, wherein the updating the privacy description mining result based on the attention index and performing personalized privacy description mining and information anonymization operation on the cloud service user interaction record by using the privacy description mining result with the updated description size comprises:

updating the description size of the privacy description mining result to enable the number of the attention indexes of the privacy description mining result to be consistent with the number of the attention indexes of the privacy distribution of the key service items recorded by the X group of cloud service user interaction;

carrying out personalized privacy description mining on service session behavior events of key service item privacy distribution recorded by the X group of cloud service user interaction by using the privacy description mining result of which the description size is updated, so as to obtain first key service item privacy distribution;

and performing information anonymization operation on the privacy distribution of the first key service item to obtain the X-th group of cloud service user interaction records for completing individual privacy anonymization processing.

3. The method according to claim 2, wherein the personalized privacy description mining is performed on the service session behavior event of the cloud service user interaction record to be subjected to user information protection, and a privacy description mining result is obtained, and the method comprises:

performing personalized privacy description mining on the cloud service user interaction record to be subjected to user information protection to determine that the service session behavior event of the Yth group of cloud service user interaction records is compared with sensitive event content change data of the service session behavior event of the Xth group of cloud service user interaction records, and obtaining related behavior event content, wherein the sensitive event content change data comprise an event heat value and an event interaction state;

and carrying out individual image recognition operation on the associated behavior event content to obtain the privacy description mining result.

4. The method of claim 3, wherein after performing personalized privacy description mining on the cloud service user interaction record to be subjected to user information protection to determine that the service session behavior event of the Yth group of cloud service user interaction records is compared with sensitive event content change data of the service session behavior event of the Xth group of cloud service user interaction records, and obtaining associated behavior event content, the method further comprises: and performing personalized privacy description mining on the service session behavior events of the key service item privacy distribution recorded by the Y-th group of cloud service user interaction by using the associated behavior event content to obtain second key service item privacy distribution.

5. The method according to claim 4, wherein the performing personalized privacy description mining on the service session behavior events of the key service item privacy distribution of the Yth group of cloud service user interaction records by using the associated behavior event content to obtain a second key service item privacy distribution comprises:

updating the description size of the associated behavior event content to enable the number of the attention indexes of the associated behavior event content to be consistent with the number of the attention indexes of the privacy distribution of the key service items of the interaction records of the Yth group of cloud service users;

carrying out personalized privacy description mining on the service session behavior events of the key service item privacy distribution of the interaction records of the cloud service users of the Yth group, which finish individual privacy anonymization processing, by using the associated behavior event content with the updated description size, so as to obtain the second key service item privacy distribution;

the performing information anonymization operation on the privacy distribution of the first key service item to obtain the X-th group of cloud service user interaction records for completing individual privacy anonymization processing comprises the following steps:

splicing the first key service item privacy distribution and the second key service item privacy distribution to obtain a third key service item privacy distribution;

and carrying out information anonymization operation on the privacy distribution of the third key service item to obtain the X group of cloud service user interaction records which are subjected to individual privacy anonymization processing.

6. The method according to claim 3, wherein the performing personalized privacy description mining on the cloud service user interaction record to be subjected to user information protection to determine that the service session behavior event of the Yth group of cloud service user interaction records is compared with the sensitive event content change data of the service session behavior event of the Xth group of cloud service user interaction records, and obtaining associated behavior event content comprises:

combining the X group of cloud service user interaction records, the Y group of cloud service user interaction records and the Y group of cloud service user interaction records which are subjected to individual privacy anonymization processing under the attention level to obtain the cloud service user interaction records to be subjected to user information protection;

performing individual portrait recognition operation on the cloud service user interaction record to be subjected to user information protection to obtain privacy distribution of a fourth key service item;

performing personalized privacy description mining on the privacy distribution of the fourth key service item to obtain the privacy distribution of a fifth key service item;

updating the attention index quantity of the privacy distribution of the fifth key service item to a first designated numerical value by utilizing personalized privacy description mining, and obtaining the content of the associated behavior event.

7. The method of claim 6, wherein the associating the behavior event content to perform an individual image recognition operation to obtain the privacy description mining result comprises:

updating the attention index quantity of the associated behavior event content to a second specified numerical value by utilizing personalized privacy description mining to obtain privacy distribution of a sixth key service item;

splicing the fourth key service item privacy distribution and the sixth key service item privacy distribution to obtain a seventh key service item privacy distribution;

performing personalized privacy description mining on the privacy distribution of the seventh key service item to determine that the service session behavior event of the cloud service user interaction record of which the Yth group completes the individual privacy anonymization processing is compared with the individual privacy data of the service session behavior event of the cloud service user interaction record of the Yth group, and obtaining the privacy description mining result;

the personalized privacy description mining of the privacy distribution of the seventh key service item to determine that the interaction record of the cloud service user completing the individual privacy anonymization processing in the Y group is compared with the individual privacy data of the service session behavior event of the interaction record of the cloud service user in the Y group, and obtaining the privacy description mining result includes:

carrying out personalized privacy description mining on the privacy distribution of the seventh key service item to obtain the privacy distribution of the eighth key service item;

updating the attention index quantity of the eighth key service item privacy distribution to the first designated numerical value by utilizing personalized privacy description mining, and obtaining the privacy description mining result.

8. The method of claim 5, wherein the performing information anonymization on the third key service item privacy distribution to obtain the xth group of cloud service user interaction records that are processed by individual privacy anonymization includes:

performing k-anonymity privacy analysis on the privacy distribution of the third key service item to obtain the privacy distribution of a ninth key service item;

performing personalized privacy description mining on the ninth key service item privacy distribution to obtain a cloud service user interaction record after anonymous operation of the Xth group of information;

weighting an event subject label of a first service session behavior event recorded by the X group of cloud service user interaction records with an event subject label of a second service session behavior event recorded by the cloud service user interaction records after the X group of information is anonymously operated to obtain the X group of cloud service user interaction records after the individual privacy anonymization processing is completed, wherein the positioning analysis data of the first service session behavior event in the X group of cloud service user interaction records is consistent with the positioning analysis data of the second service session behavior event in the X group of cloud service user interaction records after the X group of information is anonymously operated.

9. A smart cloud server, comprising: a memory and a processor; the memory and the processor are coupled; the memory for storing computer program code, the computer program code comprising computer instructions; wherein the computer instructions, when executed by the processor, cause the smart cloud server to perform the method of any of claims 1-8.