CN111814054A - Recommendation method and recommendation device for mass information data - Google Patents
Recommendation method and recommendation device for mass information data Download PDFInfo
- Publication number
- CN111814054A CN111814054A CN202010697431.3A CN202010697431A CN111814054A CN 111814054 A CN111814054 A CN 111814054A CN 202010697431 A CN202010697431 A CN 202010697431A CN 111814054 A CN111814054 A CN 111814054A
- Authority
- CN
- China
- Prior art keywords
- recommendation
- information
- application service
- engine
- unit
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9535—Search customisation based on user profiles and personalisation
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F16/00—Information retrieval; Database structures therefor; File system structures therefor
- G06F16/90—Details of database functions independent of the retrieved data types
- G06F16/95—Retrieval from the web
- G06F16/953—Querying, e.g. by the use of web search engines
- G06F16/9536—Search customisation based on social or collaborative filtering
-
- G—PHYSICS
- G06—COMPUTING OR CALCULATING; COUNTING
- G06Q—INFORMATION AND COMMUNICATION TECHNOLOGY [ICT] SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES; SYSTEMS OR METHODS SPECIALLY ADAPTED FOR ADMINISTRATIVE, COMMERCIAL, FINANCIAL, MANAGERIAL OR SUPERVISORY PURPOSES, NOT OTHERWISE PROVIDED FOR
- G06Q30/00—Commerce
- G06Q30/06—Buying, selling or leasing transactions
- G06Q30/0601—Electronic shopping [e-shopping]
- G06Q30/0631—Recommending goods or services
Landscapes
- Engineering & Computer Science (AREA)
- Databases & Information Systems (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Business, Economics & Management (AREA)
- General Physics & Mathematics (AREA)
- Accounting & Taxation (AREA)
- General Engineering & Computer Science (AREA)
- Data Mining & Analysis (AREA)
- Finance (AREA)
- Development Economics (AREA)
- Economics (AREA)
- Marketing (AREA)
- Strategic Management (AREA)
- General Business, Economics & Management (AREA)
- Management, Administration, Business Operations System, And Electronic Commerce (AREA)
- Information Retrieval, Db Structures And Fs Structures Therefor (AREA)
Abstract
The invention discloses a recommendation method of mass information data, which provides a service adaptation unit according to content attribute information of an application service unit and attribute information of a recommendation engine, wherein the service adaptation unit establishes a relationship between the application service unit and the recommendation engine; the service adaptation unit receives a recommendation request provided by the application service unit; the service adaptation unit provides a recommendation request to the recommendation engine; the recommendation engine provides recommendation information to the service adaptation unit, and the service adaptation unit provides the recommendation information to the application service unit; the application service unit for directly recommending the content is separated from the recommendation engine for generating the recommendation information, the original recommendation engine is still utilized when new content is needed to be expanded and added, the development and expansion of the recommendation engine are not needed to be re-developed, and the difficulty and the workload of the expansion of the recommendation device are reduced.
Description
Technical Field
The invention relates to a data recommendation method, in particular to a recommendation method for multi-classified mass information data.
Background
With the advent of the big data era, how to help users to quickly obtain useful information from a large amount of information becomes an important task of a plurality of merchants, and a personalized recommendation system comes into play. The personalized recommendation system guides users to find own information requirements based on mass data mining, and is widely applied to many fields.
The recommendation devices so far comprise a recommendation engine suitable for one recommended content, and for a platform based on mass information data and having a plurality of categories of recommended contents, a corresponding recommendation engine needs to be established according to the attributes of the recommended contents, and recommendation of all the categories of recommended contents cannot be realized through one recommendation engine; since the data capacity and new data of the platform based on the mass information data still increase with time, the attribute of the recommended content inevitably increases, and a new corresponding recommendation engine needs to be established, so that the development difficulty of the recommendation device is high, and the recommendation device is close to the redevelopment of the recommendation device.
Disclosure of Invention
The invention provides a recommendation method for massive information data without reestablishing a recommendation engine during expansion, and solves the technical problem that a recommendation device in the related technology is difficult to expand.
According to an aspect of the present invention, there is provided a recommendation method for mass information data, including the following steps:
providing a service adaptation unit according to the content attribute information of an application service unit and the attribute information of a recommendation engine, wherein the service adaptation unit establishes a connection between the application service unit and the recommendation engine;
the service adaptation unit receives a recommendation request provided by the application service unit;
the service adaptation unit provides a recommendation request to the recommendation engine;
the recommendation engine provides recommendation information to the service adaptation unit, and the service adaptation unit provides the recommendation information to the application service unit; the application service unit provides the content associated with the recommendation information to the user through an interactive interface;
the content attribute information of the application service unit is the attribute of the content of the service provided by the application service unit;
the attribute information of the recommendation engine is a recommendation algorithm of the recommendation engine.
By adopting the technical scheme, the invention provides a recommendation method for providing a service adaptation unit according to an application service unit, wherein the service adaptation unit is used for connecting the application service unit to a recommendation engine, the service adaptation unit is matched with the application service unit to start the corresponding recommendation engine to obtain recommendation information, and the recommendation information is sent to the service adaptation unit.
The universality of the recommendation engine is fully utilized, and only the application service unit corresponding to new content needs to be added, so that the difficulty and the workload of expansion of the recommendation device are reduced.
Further, the establishment of the connection between the service adaptation unit and the recommendation engine is that the service adaptation unit connects to an application programming interface of the recommendation engine.
As another aspect of the present invention, there is provided an apparatus for recommending massive information data, including:
the big data service platform comprises a plurality of application service units and service adaptation units which are in one-to-one correspondence with the application service units;
a recommendation service platform comprising a plurality of recommendation engines;
the database platform is used for storing the data required to be called by the recommendation engine;
the method comprises the steps that a plurality of application service units respectively provide or sell contents, and each application service unit respectively provides or sells at least one content with certain attributes;
the service adaptation unit connects the corresponding application service unit with at least one recommendation engine, the service adaptation unit starts the recommendation engine based on the request of the application service unit, the recommendation engine calls data of a database platform and processes the data to obtain recommendation information, and the service adaptation unit provides the recommendation information obtained by the recommendation engine to the corresponding application service unit.
By adopting the technical scheme, the invention provides the recommendation device which provides the service adaptation unit according to the application service unit, the service adaptation unit connects the application service unit to the recommendation engine, the service adaptation unit is matched with the application service unit to start the corresponding recommendation engine to obtain the recommendation information, and the recommendation information is sent to the service adaptation unit.
Further, the plurality of recommendation engines includes at least one of:
a collaborative filtering engine based on a collaborative filtering algorithm, a probability engine based on a graph model, and a rule engine based on association rules.
Further, the service adaptation unit comprises an attribute management part and a contact part, wherein the attribute management part is used for setting content attribute information, recommendation engine information, parameter information required by the start of the recommendation engine and table information required by the start of the recommendation engine;
the attribute management part provides an interactive interface which can change the content attribute information, the recommendation engine information, the parameter information required by the start of the recommendation engine and the table information required by the start of the recommendation engine, and the user changes the content attribute information, the recommendation engine information, the parameter information required by the start of the recommendation engine and the table information required by the start of the recommendation engine of the attribute management part through the interactive interface;
the contact part calls the information of the attribute management part and starts a recommendation engine together with the application service unit corresponding to the service adaptation unit.
Further, the contact part contacts the application service unit and the recommendation engine corresponding to the service adaptation unit where the contact part is located, the contact part is in contact with an application programming interface of the recommendation engine, and the contact part further has a function of transmitting data between the recommendation engine and the application service unit.
Further, the content respectively provided by the plurality of application service units comprises patent content, expert content, result content and policy content;
the plurality of application service units respectively contact a collaborative filtering engine based on a collaborative filtering algorithm and/or a rule engine based on an association rule.
Further, the recommendation method of the collaborative filtering engine based on the collaborative filtering algorithm comprises the following steps:
extracting a feature vector and a feature-item correlation matrix of a user;
processing the characteristic-article correlation matrix, deleting articles which do not belong to the content provided by the application service unit in the characteristic-article correlation matrix, and obtaining a processed characteristic-article correlation matrix;
and obtaining recommendation information based on the feature vector of the user and the processed feature-article correlation matrix.
Further, an application service unit provides patent content, which contacts two recommendation engines, wherein the two recommendation engines are collaborative filtering engines based on a collaborative filtering algorithm;
one recommendation engine deletes the articles which do not belong to the patent content in the feature-article correlation matrix to obtain a processed feature-patent correlation matrix; obtaining patent recommendation information based on the feature vector of the user and the processed feature-patent correlation matrix;
the other recommendation engine deletes the articles which do not belong to the expert content in the feature-article correlation matrix to obtain a processed feature-expert correlation matrix; obtaining expert recommendation information based on the feature vector of the user and the processed feature-expert correlation matrix;
the patent recommendation information is screened out to be used as final recommendation information, wherein patent item sets with the first screening items and any one second screening item which are the same are used as the final recommendation information, and the application service unit provides patent contents based on the final recommendation information.
Further, the method for adding the application service unit to the big data service platform comprises the following steps:
adding an application service unit, and providing a service adaptation unit according to the content attribute information of the application service unit and the attribute information of a recommendation engine;
and contacting said service adaptation unit with at least one of said recommendation engines.
The invention has the beneficial effects that: the application service unit for directly recommending the content is separated from the recommendation engine for generating the recommendation information, the original recommendation engine is still utilized when new content is needed to be expanded and added, the development and expansion of the recommendation engine are not needed to be re-developed, and the difficulty and the workload of the expansion of the recommendation device are reduced.
Drawings
Fig. 1 is a schematic structural diagram of a recommendation apparatus for mass information data according to an embodiment of the present invention;
fig. 2 is a schematic diagram of a recommendation apparatus for mass information data according to an embodiment of the present invention, adding a new application service unit;
FIG. 3 is a schematic structural diagram of a service adaptation unit according to an embodiment of the present invention;
FIG. 4 is a data set diagram of an embodiment of the invention;
FIG. 5 is a schematic diagram of a bipartite graph model according to an embodiment of the invention;
FIG. 6 is a flowchart of a method for recommending mass information data according to an embodiment of the present invention;
fig. 7 is a flowchart of a recommendation method of a collaborative filtering engine based on a collaborative filtering algorithm according to an embodiment of the present invention.
Detailed Description
The subject matter described herein will now be discussed with reference to example embodiments. It should be understood that these embodiments are discussed only to enable those skilled in the art to better understand and thereby implement the subject matter described herein, and are not intended to limit the scope, applicability, or examples set forth in the claims. Changes may be made in the function and arrangement of elements discussed without departing from the scope of the disclosure. Various examples may omit, substitute, or add various procedures or components as needed. For example, the described methods may be performed in an order different from that described, and various steps may be added, omitted, or combined. In addition, features described with respect to some examples may also be combined in other examples.
As shown in fig. 1, a recommendation apparatus for mass information data includes a big data service platform, a recommendation service platform and a database platform,
the big data service platform comprises application service units 210, 220, 230 and 240 and service adaptation units 310, 320, 330 and 340 corresponding to the application service units in a one-to-one mode;
the recommendation service platform includes a plurality of recommendation engines 410, 420, 430;
the database platform stores data that the recommendation engines 410, 420, 430 need to invoke.
The application service units 210, 220, 230, 240 respectively provide or sell contents, each application service unit 210, 220, 230, 240 respectively provides or sells at least one or more contents with certain attributes; (ii) a
Because the content attributes of the application service units 210, 220, 230, and 240 are different, the recommendation engines are also different, so that the recommendation device contacts the big data service platform and the recommendation service platform to provide recommendation information for the big data service platform, and establishes a plurality of contact parts between the recommendation service platform and the application service units 210, 220, 230, and 240 at one time to support the content services of the application service units 210, 220, 230, and 240, and separates the contact parts from the recommendation service platform;
namely, the recommending device provides a service adapting unit for the recommending engine to contact with the application service unit.
Even if a new application service unit is added for adding a new content service, the architecture of the recommendation service platform does not need to be changed, a new recommendation engine does not need to be added, and only a service adaptation unit corresponding to the new application service unit needs to be added.
As shown in fig. 2, the application service unit 250 is added, and only the service adaptation unit 350 corresponding to the application service unit 250 needs to be added and is associated with at least one recommendation engine 410, 420, 430.
The recommendation engines 410, 420, 430 are implemented based on a variety of recommendation algorithms, which may be selected but not limited to: a collaborative filtering engine based on a collaborative filtering algorithm, a probability engine based on a graph model, and a rule engine based on an association rule. The recommendations of the recommendation engines 410, 420, 430 are based on data of the database platform;
the collaborative filtering engine adopts a user-based collaborative filtering algorithm (user-based collaborative filtering), namely a UserCF algorithm, and identifies similar users with similar preference and inclination on the basis of the preference and inclination of the user A. The collaborative filtering engine recommends items similar to the items selected by the user, which are not yet selected by the user A, and selects related items according to the preference or life form of the user A through cross recommendation or classification.
The aforementioned selection can be various embodied modes of purchase, scoring, browsing, and the like.
The article may be embodied in various forms such as data, information, a commodity, and the like.
The probability engine adopts a recommendation algorithm based on a graph model, reads the bipartite graph model from the database platform, and recommends the articles for the user by using a stochastic migration personalRank algorithm:
assuming that user u is to be recommended individually, node u corresponding to user u may be selecteduStarting random walk on the user item bipartite graph, and when the user item moves to any node, determining whether to continue the walk or stop the walk according to the probability of alpha and starting from uuAnd the nodes are walked again, if the nodes are continuously walked, one node is randomly selected from the nodes pointed by the current node as the next node of the walking according to uniform distribution.
After repeating the above steps, the probability of each item node being accessed converges to a number, the weight of the item in the final recommendation list is the access probability of the item node, and TopN items are recommended.
The bipartite graph model is established based on user behavior data, the user behavior data are expressed as a series of binary groups, each binary group (u, i) represents that a user u generates behavior on an article i, and the data set is expressed as a bipartite graph model;
assuming a data set as shown in fig. 4, only the user likes and dislikes the item and does not consider the user's likeness to the item, where user is ═ a, B, C, and item is ═ a, B, C, d, data and a bipartite graph model as shown in fig. 5 are represented by G (V, E), where V is a vertex set consisting of user and item, i.e., [ a, B, C, d ], and E represents a corresponding edge E (u, i) between each two-tuple (u, i).
The rule engine adopts a recommendation algorithm based on association rules, acquires user historical access, scores or purchase records and the like, and then recommends the articles to the user according to the association rules among the articles.
The association rule between the articles is stored in a database platform and can be directly called;
the contents provided or sold by the application service units 210, 220, 230, 240 include user technical contents, expert guidance contents, achievement transformation contents, information contents of government policies, music contents, video contents, avatar contents, movie contents, book contents, applications, game contents, news contents, advertisement contents, regional basic contents.
The service adaptation units 310, 320, 330, 340 correspond to the application service units 210, 220, 230, 240, respectively, and the service adaptation units 310, 320, 330, 340 provide recommendation information for the corresponding application service units 210, 220, 230, 240. The service adaptation units 310, 320, 330, 340 contact the corresponding application service units 210, 220, 230, 240 with the recommendation engine. If the application service units are expanded, corresponding service adaptation units are added according to the expanded application service units, and the service adaptation units can be generated by a big data service platform.
The service adaptation unit 310, 320, 330, 340 selects a recommendation engine on the basis of the content property information of the corresponding application service unit 210, 220, 230, 240.
The service adaptation units 310, 320, 330, 340 contact the application service units 210, 220, 230, 240 with at least one more recommendation engines 410, 420, 430, respectively.
For example, if the application service unit 210 wants to obtain recommendation information from the recommendation engine 410, the server assembly 310 obtains the recommendation information from the recommendation engine 410 and provides the application service unit 210 with the recommendation information.
If the application service unit 220 wants to obtain the recommendation information from the recommendation engines 410, 420, the server assembly 320 will obtain the recommendation information from the recommendation engines 410, 420 and provide it to the application service unit 220 to contact between the application service unit 220 and the recommendation engines 410, 420.
If the application service unit 230 wants to obtain recommendation information from the recommendation engines 410, 420, 430, the server assembly 330 will obtain the recommendation information from the recommendation engines 410, 420, 430 and provide it to the application service unit 230 to contact between the application service unit 230 and the recommendation engines 410, 420, 430.
If the application service unit 240 wants to obtain recommendation information from the recommendation engines 410, 420, 430, the server assembly 340 will obtain the recommendation information from the recommendation engines 410, 420, 430 and provide it to the application service unit 240 to contact between the application service unit 240 and the recommendation engines 410, 420, 430.
The database platform stores data that the recommendation engines 410, 420, 430 need to invoke. Such as user information, situation information, record information. The user information may include user personal information, user tendency information, and the like;
the situation information may include time-date, weather, and location information;
the record information may include various record information such as a content usage record, a browsing record, a search record, and the like.
The device can expand or delete the application service units according to the content requirements, and adds the corresponding service adaptation units to be connected to the recommendation engines 410, 420 and 430 of the recommendation service platform, so that the recommendation engines 410, 420 and 430 do not need to be changed, the expansion can be simpler, and the content can be adapted to the mass information data which changes constantly.
Referring to fig. 3, the service adaptation unit 310, 320, 330, 340 should include at least an attribute management part 301 and a contact part 302, the attribute management part 301 being used to set content attribute information, recommendation engine information, parameter information required for the start of a recommendation engine, and various table information required for the start of a recommendation engine. The content attribute information may include the type of content such as movies, music, and books as an attribute of the content, which is an attribute of the content that the application service unit provides the service. An optimal recommendation engine may be selected based on the content attribute information.
The recommendation attribute information includes various recommendation conditions, such as a recommendation period, a learning period, and the like, for recommending to the application service unit.
The recommendation engine information is recommendation engine information configured at the application service unit.
The attribute management unit 301 provides an interactive interface that can change the above information, and the user can change various information of the attribute management unit 301 by accessing the interactive interface.
The contact unit 302 calls the information of the attribute management unit 301 and starts the recommendation engine together with the corresponding application service unit.
The contact part 302 contacts the corresponding application service unit and the recommendation engine, the contact part 302 is in contact with an application programming interface of the recommendation engine, and the contact part 302 also has a function of transmitting data between the recommendation engine and the application service unit.
There are various selectable attributes in the content with certain attributes of the application service units 210, 220, 230, and 240, and the attributes can be adaptively selected according to a specific application field, and two selectable modes are provided in this embodiment:
example one
The big data service platform is used as an information providing and service platform for providing user technology, expert guidance, achievement transformation and government policy;
the application service unit 210 provides patent content, the application service unit 220 provides expert content, the application service unit 230 provides achievement content, and the application service unit 240 provides policy content;
for example, according to the attribute (category of content) of the patent content of the application service unit 210, the collaborative filtering engine based on the collaborative filtering algorithm and/or the rule engine based on the association rule are selected according to the attribute of the patent content.
Such a recommendation device can provide recommendations of patent content, expert content, achievement content and policy content for enterprise users, and assist the enterprise users in better development.
For the expansion of the application service unit, the expanded application service unit 250 provides the content of the developer. Thereby increasing recommendations for developer content.
As shown in fig. 7, the recommendation method of the collaborative filtering engine based on the collaborative filtering algorithm includes:
extracting a feature vector and a feature-item correlation matrix of a user;
processing the characteristic-article correlation matrix, deleting articles which do not belong to the content provided by the application service unit in the characteristic-article correlation matrix, and obtaining a processed characteristic-article correlation matrix;
for example, for the application service unit 210, items in the feature-item correlation matrix that are not proprietary are deleted.
For the application service unit 220, items in the feature-item correlation matrix that do not belong to expert content are deleted.
For the application service unit 230, items in the feature-item correlation matrix that do not belong to the outcome content are deleted.
For the application service unit 240, items in the feature-item correlation matrix that do not belong to the policy content are deleted.
Obtaining recommendation information based on the feature vector of the user and the processed feature-article correlation matrix;
the recommendation information should be a sorted item list (patent example contains several patent items), and further, the final recommendation information can be obtained through filtering and/or ranking.
For example, the application service unit 210 provides patent content, which contacts the recommendation engines 410, 420, and the recommendation engine 410 and the recommendation engine 420 are collaborative filtering engines based on collaborative filtering algorithms;
the recommendation engine 410 deletes the articles in the feature-article correlation matrix that do not belong to the patent content to obtain a processed feature-patent correlation matrix; obtaining patent recommendation information based on the feature vector of the user and the processed feature-patent correlation matrix;
the recommendation engine 420 deletes the items in the feature-item correlation matrix that do not belong to the expert content to obtain a processed feature-expert correlation matrix; obtaining expert recommendation information based on the feature vector of the user and the processed feature-expert correlation matrix;
the inventor name of a patent item in the patent recommendation information is used as a first screening item, the expert name of an expert item of the expert recommendation information is used as a second screening item, and a patent item set with the first screening item being the same as any one of the second screening items is screened out from the patent recommendation information of the recommendation engine 410 to serve as final recommendation information.
The screening process may be performed in the application service unit 210.
The recommending device and the recommending method based on the recommending device can provide more accurate patent content with less total amount for the user.
The content provided by the application service unit is not limited to the above-mentioned one, but the application service unit 210 may provide patent content, the application service unit 220 provides expert content, the application service unit 230 provides outcome content and patent content, and the application service unit 240 provides policy content; and other content selections provided by the application service unit.
Example two
The big data service platform is applied to the technical field of commerce and is used as an information providing and service platform for providing commodities, advertisements, images and music;
the application service unit 210 provides commodity contents, the application service unit 220 provides advertisement contents, the application service unit 230 provides image contents, and the application service unit 240 provides music contents;
or the application service unit 210 provides commodity contents, the application service unit 220 provides advertisement contents, the application service unit 230 provides image contents and music contents, and the application service unit 240 provides music contents;
for example, according to the attribute (category of content) of the patent content of the application service unit 220, the collaborative filtering engine based on the collaborative filtering algorithm and/or the probability engine based on the graph model are selected according to the attribute of the patent content.
The recommendation device can provide recommendation of goods, advertisements, images and music contents for enterprise users, and provides good experience for shopping of the users.
For the expansion of the application service unit, the expanded application service unit 250 provides news content. Thereby increasing the recommendation of news content.
As shown in fig. 6, a recommendation method for mass information data includes the following steps:
s100, providing a service adaptation unit according to the content attribute information of the application service unit and the attribute information of the recommendation engine, wherein the service adaptation unit establishes a connection between the application service unit and the recommendation engine;
s200, the service adaptation unit receives a recommendation request provided by the application service unit;
s300, the service adaptation unit provides a recommendation request to a recommendation engine;
s400, the recommendation engine provides recommendation information to a service adaptation unit, and the service adaptation unit provides the recommendation information to an application service unit;
the application service unit provides the content associated with the recommendation information, such as patents in the recommendation information, the application service unit provides the associated patent content (including text and the like), music in the recommendation information, and the application service unit provides the associated music content (including text and the like), which can be read from the database platform, to the user through the interactive interface.
The establishment of the connection between the service adaptation unit and the recommendation engine is that the service adaptation unit connects the application programming interface of the recommendation engine.
The service adaptation unit provides the recommendation information to the application service unit based on the related information of the application service unit, and the related information of the application service unit comprises at least one of project file information, situation information and use detail information related to the application service unit. I.e. establishes contact with the application service unit based on the relevant information of the application service unit.
The content attribute information of the application service unit is an attribute of the content of the service provided by the application service unit, and may include the category of the content such as movie, music, book, etc. as the attribute of the content. An optimal recommendation engine may be selected based on the content attribute information.
The attribute information of the recommendation engine, that is, the recommendation algorithm of the recommendation engine, may be selected but not limited to: a collaborative filtering engine based on a collaborative filtering algorithm, a probability engine based on a graph model, and a rule engine based on an association rule.
The service adaptation unit in S100 is provided in the embodiment specifically in the following two ways:
the plurality of service adaptation units can be pre-stored on a big data service platform and provided in a calling mode.
And secondly, the service adaptation unit is generated by the big data service platform according to the existing generation method according to the content attribute information of the application service unit and the attribute information of the recommendation engine.
Claims (10)
1. A recommendation method for massive information data is characterized by comprising the following steps:
providing a service adaptation unit according to the content attribute information of an application service unit and the attribute information of a recommendation engine, wherein the service adaptation unit establishes a connection between the application service unit and the recommendation engine;
the service adaptation unit receives a recommendation request provided by the application service unit;
the service adaptation unit provides a recommendation request to the recommendation engine;
the recommendation engine provides recommendation information to the service adaptation unit, and the service adaptation unit provides the recommendation information to the application service unit; the application service unit provides the content associated with the recommendation information to the user through an interactive interface;
the content attribute information of the application service unit is the attribute of the content of the service provided by the application service unit;
the attribute information of the recommendation engine is a recommendation algorithm of the recommendation engine.
2. The method for recommending massive information data according to claim 1, wherein the establishing of the association between the service adaptation unit and the recommendation engine is the service adaptation unit associating with an application programming interface of the recommendation engine.
3. The mass information data recommendation device according to claim, comprising:
the big data service platform comprises a plurality of application service units and service adaptation units which are in one-to-one correspondence with the application service units;
a recommendation service platform comprising a plurality of recommendation engines;
the database platform is used for storing the data required to be called by the recommendation engine;
the method comprises the steps that a plurality of application service units respectively provide or sell contents, and each application service unit respectively provides or sells at least one content with certain attributes;
the service adaptation unit connects the corresponding application service unit with at least one recommendation engine, the service adaptation unit starts the recommendation engine based on the request of the application service unit, the recommendation engine calls data of a database platform and processes the data to obtain recommendation information, and the service adaptation unit provides the recommendation information obtained by the recommendation engine to the corresponding application service unit.
4. The apparatus for recommending massive information data according to claim 3, wherein said plurality of recommendation engines comprises at least one of the following recommendation engines:
a collaborative filtering engine based on a collaborative filtering algorithm, a probability engine based on a graph model, and a rule engine based on association rules.
5. The massive information data recommending device according to claim 3 or 4, wherein said service adapting unit comprises an attribute managing part and a contact part, said attribute managing part is used for setting content attribute information, recommendation engine information, parameter information required for recommending engine startup, and table information required for recommending engine startup;
the attribute management part provides an interactive interface which can change the content attribute information, the recommendation engine information, the parameter information required by the start of the recommendation engine and the table information required by the start of the recommendation engine, and the user changes the content attribute information, the recommendation engine information, the parameter information required by the start of the recommendation engine and the table information required by the start of the recommendation engine of the attribute management part through the interactive interface;
the contact part calls the information of the attribute management part and starts a recommendation engine together with the application service unit corresponding to the service adaptation unit.
6. The massive information data recommendation device according to claim 5, wherein the contact part contacts the application service unit and the recommendation engine corresponding to the service adaptation unit where the contact part is located, the contact part is in contact with an application programming interface of the recommendation engine, and the contact part further has a function of transmitting data between the recommendation engine and the application service unit.
7. The apparatus for recommending massive information data according to claim 3, wherein the contents provided by the plurality of application service units respectively include patent contents, expert contents, achievement contents and policy contents;
the plurality of application service units respectively contact a collaborative filtering engine based on a collaborative filtering algorithm and/or a rule engine based on an association rule.
8. The mass information data recommendation device according to claim 3 or 7, wherein the recommendation method of the collaborative filtering engine based on the collaborative filtering algorithm comprises:
extracting a feature vector and a feature-item correlation matrix of a user;
processing the characteristic-article correlation matrix, deleting articles which do not belong to the content provided by the application service unit in the characteristic-article correlation matrix, and obtaining a processed characteristic-article correlation matrix;
and obtaining recommendation information based on the feature vector of the user and the processed feature-article correlation matrix.
9. The mass information data recommendation device according to claim 8, wherein an application service unit provides patent content, which contacts two recommendation engines, both of which are collaborative filtering engines based on collaborative filtering algorithm;
one recommendation engine deletes the articles which do not belong to the patent content in the feature-article correlation matrix to obtain a processed feature-patent correlation matrix; obtaining patent recommendation information based on the feature vector of the user and the processed feature-patent correlation matrix;
the other recommendation engine deletes the articles which do not belong to the expert content in the feature-article correlation matrix to obtain a processed feature-expert correlation matrix; obtaining expert recommendation information based on the feature vector of the user and the processed feature-expert correlation matrix;
the patent recommendation information is screened out to be used as final recommendation information, wherein patent item sets with the first screening items and any one second screening item which are the same are used as the final recommendation information, and the application service unit provides patent contents based on the final recommendation information.
10. The mass information data recommendation device according to claim 3, wherein the method for adding the application service unit to the big data service platform comprises:
adding an application service unit, and providing a service adaptation unit according to the content attribute information of the application service unit and the attribute information of a recommendation engine;
and contacting said service adaptation unit with at least one of said recommendation engines.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010697431.3A CN111814054B (en) | 2020-07-20 | 2020-07-20 | Recommendation method and recommendation device for mass information data |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010697431.3A CN111814054B (en) | 2020-07-20 | 2020-07-20 | Recommendation method and recommendation device for mass information data |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN111814054A true CN111814054A (en) | 2020-10-23 |
| CN111814054B CN111814054B (en) | 2021-11-09 |
Family
ID=72865556
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010697431.3A Active CN111814054B (en) | 2020-07-20 | 2020-07-20 | Recommendation method and recommendation device for mass information data |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN111814054B (en) |
Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6889222B1 (en) * | 2000-12-26 | 2005-05-03 | Aspect Communications Corporation | Method and an apparatus for providing personalized service |
| US20120054065A1 (en) * | 2010-08-24 | 2012-03-01 | Electronics And Telecommunications Research Institute | System and method for service recommendation |
| US20130159325A1 (en) * | 2010-02-03 | 2013-06-20 | Gartner, Inc. | Bi-model recommendation engine for recommending items and peers |
| US20140257890A1 (en) * | 2013-03-05 | 2014-09-11 | Gartner, Inc. | Methods and systems for improving engagement with a recommendation engine that recommends items, peers, and services |
| CN106600302A (en) * | 2015-10-19 | 2017-04-26 | 玺阅信息科技（上海）有限公司 | Hadoop-based commodity recommendation system |
| CN107357845A (en) * | 2017-06-26 | 2017-11-17 | 西安理工大学 | A kind of tour interest commending system and recommendation method based on Spark |
| CN109670116A (en) * | 2018-11-30 | 2019-04-23 | 内江亿橙网络科技有限公司 | A kind of intelligent recommendation system based on big data |
-
2020
- 2020-07-20 CN CN202010697431.3A patent/CN111814054B/en active Active
Patent Citations (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US6889222B1 (en) * | 2000-12-26 | 2005-05-03 | Aspect Communications Corporation | Method and an apparatus for providing personalized service |
| US20130159325A1 (en) * | 2010-02-03 | 2013-06-20 | Gartner, Inc. | Bi-model recommendation engine for recommending items and peers |
| US20120054065A1 (en) * | 2010-08-24 | 2012-03-01 | Electronics And Telecommunications Research Institute | System and method for service recommendation |
| US20140257890A1 (en) * | 2013-03-05 | 2014-09-11 | Gartner, Inc. | Methods and systems for improving engagement with a recommendation engine that recommends items, peers, and services |
| CN106600302A (en) * | 2015-10-19 | 2017-04-26 | 玺阅信息科技（上海）有限公司 | Hadoop-based commodity recommendation system |
| CN107357845A (en) * | 2017-06-26 | 2017-11-17 | 西安理工大学 | A kind of tour interest commending system and recommendation method based on Spark |
| CN109670116A (en) * | 2018-11-30 | 2019-04-23 | 内江亿橙网络科技有限公司 | A kind of intelligent recommendation system based on big data |
Non-Patent Citations (2)
| Title |
|---|
| 刘伟等: "海量学术资源个性化推荐综述", 《计算机工程与应用》 * |
| 百分点科技推荐引擎: "百分点推荐引擎——从需求到架构", 《HTTPS://WWW.INFOQ.CN/ARTICLE/BAIFENDIAN-RECOMMENDATION-ENGINE》 * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN111814054B (en) | 2021-11-09 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| US11875363B2 (en) | Method and system for generating recommendations | |
| US7693887B2 (en) | Dynamic identification of a new set of media items responsive to an input mediaset | |
| US6438579B1 (en) | Automated content and collaboration-based system and methods for determining and providing content recommendations | |
| US8782035B2 (en) | Syndicated data stream content provisioning | |
| US6092049A (en) | Method and apparatus for efficiently recommending items using automated collaborative filtering and feature-guided automated collaborative filtering | |
| US9946761B2 (en) | Recommendation for diverse content | |
| US8019707B2 (en) | Interactive hybrid recommender system | |
| US6049777A (en) | Computer-implemented collaborative filtering based method for recommending an item to a user | |
| Kamahara et al. | A community-based recommendation system to reveal unexpected interests | |
| US20090100094A1 (en) | Recommendation system and method for multimedia content | |
| JP2002520739A (en) | Systems, methods and products that increase user value recommendations | |
| CN100357941C (en) | Intelligent searching system and method for product type record | |
| JP2002520743A (en) | System, method and product for high user value recommendation | |
| KR20110133024A (en) | Single, co- and auto-generated media guides for online content | |
| US20070136247A1 (en) | Computer-implemented system and method for obtaining customized information related to media content | |
| EP1906316A1 (en) | An interactive hybrid recommender system | |
| US20070282874A1 (en) | Content System | |
| WO2001037193A1 (en) | System, method, and article of manufacture for recommending items to users based on user preferences | |
| CN111814054B (en) | Recommendation method and recommendation device for mass information data | |
| TWI468956B (en) | Method and system for personalizedly sorting searched information | |
| Schafer et al. | View through MetaLens: usage patterns for a meta-recommendation system | |
| Kar et al. | Knowledge-Based, Ensemble-Based, and Hybrid Recommender Systems | |
| HK1192338A (en) | Method and system for generating recommendations | |
| HK1116558A (en) | Method and system for generating recommendations | |
| HK1192338B (en) | Method and system for generating recommendations |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |