CN117009528A - Business processing method, device, equipment and medium based on natural language processing - Google Patents

Abstract

The application discloses a business processing method, a device, equipment and a medium based on natural language processing, which can be applied to the artificial intelligence field or the financial field, and the method comprises the following steps: and acquiring a service processing request input by a user, classifying the service processing request by using a text classification model, and determining the target type service. And carrying out entity identification on the service processing request to obtain an entity identification result. And matching the entity identification result with a plurality of problem templates in the target database, and determining a target problem matched with the entity identification result, so that a processing result corresponding to the target problem is sent to a user. The natural language processing method comprises text classification, entity identification and the like, can automatically process the service processing request of the user, and obtain corresponding processing results in the database of the corresponding service, does not need manual processing by staff, reduces the service processing time of the user, and improves the service processing efficiency.

Description

Business processing methods, devices, equipment and media based on natural language processing

Technical field

This application relates to the field of artificial intelligence technology, and in particular to a business processing method, device, equipment and medium based on natural language processing.

Background technique

With the development of economic and technological construction, the business of banks has become more and more diversified. Currently, when users handle banking services, they usually need to go to a manual window. However, with the diversification of banking business, the scope and knowledge involved have become wider and wider, and sometimes staff cannot solve users' business problems quickly and accurately. When staff review relevant information and then help users process it, it may take a lot of time and affect users' business processing.

Contents of the invention

In view of this, this application provides a business processing method, device, equipment and medium based on natural language processing, so that processing can be automatically performed according to user requests and reduce the time for users to handle business.

In the first aspect, this application provides a business processing method based on natural language processing. The method includes:

Obtain the business processing request input by the user;

Use a text classification model to classify the business processing request and determine the target type of business. The text classification model is trained based on a text collection and a business label corresponding to the text collection;

Perform entity identification on the business processing request and obtain the entity identification result;

Match the entity recognition results with multiple problem templates in a target database to determine the target problem. The target database includes multiple problem templates and processing results corresponding to each problem template. The target database is consistent with the target Corresponding to the type of business;

Send the processing results corresponding to the target problem to the user.

In a possible implementation, the training process of the text classification model includes:

Obtain a training sample set, the training sample set includes a text set, and each text in the text set includes a business label;

Taking the text collection as input and the business tags respectively corresponding to the text collection as desired output, an initial classification model is trained to obtain the text classification model.

In a possible implementation, the text collection is used as input, and the business labels corresponding to the text collection are used as desired outputs, and an initial classification model is trained to obtain the text classification model, including:

Input the text collection into the initial classification model and obtain the prediction label corresponding to the text collection;

Determine the loss function according to the business labels corresponding to the prediction labels and the text collection respectively;

When the loss function is greater than the preset value, adjust the parameters of the initial classification model, and re-execute inputting the text collection into the initial classification model and the subsequent training process until the loss function is less than or equal to the preset value. value to obtain the text classification model.

In a possible implementation, matching the entity recognition result with a question template in the target database to determine the target question includes:

Calculate the similarity between the entity recognition result and the multiple question templates respectively to obtain multiple similarity indicators;

The question template corresponding to the maximum value among the plurality of similarity indicators is determined to be the target question.

In a possible implementation, the obtaining the business processing request input by the user includes:

Get the initial processing request input by the user;

Preprocess the initial processing request to obtain the service processing request.

In a second aspect, this application provides a business processing device based on natural language processing. The device includes:

The acquisition unit is used to obtain the business processing request input by the user;

A classification unit, configured to classify the business processing request using a text classification model and determine the target type of business. The text classification model is trained based on a text collection and a business label corresponding to the text collection;

An identification unit, used to perform entity identification on the business processing request and obtain the entity identification result;

A matching unit, used to match the entity recognition result with multiple question templates in a target database to determine the target question. The target database includes multiple question templates and processing results corresponding to each question template. The target The database corresponds to the target type of business;

A sending unit, configured to send the processing results corresponding to the target question to the user.

In a possible implementation, the training process of the text classification model includes:

Obtain a training sample set, the training sample set includes a text set, and each text in the text set includes a business label; use the text set as input, and the business labels corresponding to the text set as the desired output, Train an initial classification model to obtain the text classification model.

In a possible implementation, the process of taking the text collection as input, and the business labels corresponding to the text collection as the desired output, and training an initial classification model to obtain the text classification model includes:

Input the text set into the initial classification model to obtain the prediction label corresponding to the text collection; determine the loss function according to the prediction label and the business label corresponding to the text collection; when the loss function is greater than When the preset value is used, adjust the parameters of the initial classification model, and re-execute inputting the text collection into the initial classification model, and the subsequent training process until the loss function is less than or equal to the preset value, and obtain the text Classification model.

In a possible implementation, the matching unit is specifically configured to calculate the similarity between the entity recognition result and the multiple question templates respectively to obtain multiple similarity indicators; determine the multiple similarities The problem template corresponding to the maximum value in the degree index is the target problem.

In a possible implementation, the obtaining unit is specifically configured to obtain an initial processing request input by a user; preprocess the initial processing request to obtain the business processing request.

In a third aspect, this application provides a business processing device based on natural language processing, where the device includes: a memory and a processor;

The memory is used to store relevant program codes;

The processor is configured to call the program code to execute the business processing method based on natural language processing described in any implementation of the first aspect.

In a fourth aspect, the present application provides a computer-readable storage medium, the computer-readable storage medium being used to store a computer program, the computer program being used to execute the natural-based method described in any one of the above-mentioned implementations of the first aspect. A business approach to language processing.

It can be seen that this application has the following beneficial effects:

In the above implementation manner of the present application, in order to process the user's request, after obtaining the business processing request input by the user, the text classification model can be used to classify the business processing request and determine the target type of business to which the request belongs. Among them, the text classification model is trained based on the text collection and the business tags corresponding to the text collection. And perform entity identification on the business processing request and obtain the entity identification results. According to the target type of business corresponding to the user request, a target database corresponding to the target type of business is determined, where the target database includes multiple question templates and processing results corresponding to each question template. The entity recognition result is matched with multiple question templates in the target database, and the target question matching the entity recognition result is determined, so that the processing result corresponding to the target question can be sent to the user. By using natural language processing methods, including text classification and entity recognition, users' business processing requests can be automatically processed, and the processing results corresponding to the processed requests can be obtained in the corresponding business database, without the need for manual processing by staff. , reducing the time for users to handle business, thereby improving the efficiency of handling business.

Description of the drawings

In order to more clearly illustrate the technical solutions in the embodiments of the present application, the drawings needed to be used in the description of the embodiments will be briefly introduced below. Obviously, the drawings in the following description are only some of the embodiments provided in the present application. , for those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

Figure 1 is a flow chart of a business processing method based on natural language processing provided by an embodiment of the present application;

Figure 2 is a schematic diagram of a business processing device based on natural language processing provided by an embodiment of the present application;

FIG. 3 is a schematic diagram of a business processing device based on natural language processing provided by an embodiment of the present application.

Detailed ways

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. The described embodiments are only exemplary implementations of the present application and are not all implementations. Those skilled in the art can combine the embodiments of the present application to obtain other embodiments without performing creative work, and these embodiments are also within the protection scope of the present application.

It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all It is information and data authorized by the user or fully authorized by all parties, and the collection, use and processing of relevant data need to comply with the relevant laws, regulations and standards of relevant countries and regions.

With the development of economic and technological construction, the business of banks has become more and more diversified. Currently, when users handle banking services, they usually need to go to a manual window. However, with the diversification of banking business, the scope and knowledge involved have become wider and wider, and sometimes staff cannot solve users' business problems quickly and accurately. When staff review relevant information and then help users process it, it may take a lot of time and affect users' business processing.

Based on this, embodiments of the present application provide a business processing method based on natural language processing, so that processing can be automatically performed according to user requests and reduce the time for users to handle business. In specific implementation, after obtaining the business processing request input by the user, the text classification model can be used to classify the business processing request and determine the target type of business to which the request belongs. Among them, the text classification model is trained based on the text collection and the business tags corresponding to the text collection. And perform entity identification on the business processing request and obtain the entity identification results. According to the target type of business corresponding to the user request, a target database corresponding to the target type of business is determined, where the target database includes multiple question templates and processing results corresponding to each question template. The entity recognition result is matched with multiple question templates in the target database, and the target question matching the entity recognition result is determined, so that the processing result corresponding to the target question can be sent to the user. By using natural language processing methods, including text classification and entity recognition, users' business processing requests can be automatically processed, and the processing results corresponding to the processed requests can be obtained in the corresponding business database, without the need for manual processing by staff. , reducing the time for users to handle business, thereby improving the efficiency of handling business.

In order to facilitate understanding of the technical solutions provided by the embodiments of the present application, a detailed introduction will be made below with reference to the accompanying drawings.

Referring to Figure 1, Figure 1 is a flow chart of a business processing method based on natural language processing provided by an embodiment of the present application.

The method can be executed by a data processing device, which can be a terminal or a server. Among them, terminals include but are not limited to desktop computers, laptop computers, tablet computers and smartphones. The server can be a cloud server in a cloud environment or a server in a local data center. It should also be noted that the server can be a single server or a server cluster formed by multiple servers.

The method may include the following steps:

S101: Obtain the business processing request input by the user.

When users handle business in the bank, they can input business processing requests to the data processing equipment. For example, the business processing request can be input by text on the input interface of the data processing device, or the business processing request can be input by voice through voice dialogue. As an example, the data processing device may be an intelligent robot, used to intelligently process the user's business processing request.

In a possible implementation, the business processing request input by the user may include information irrelevant to the actual business processing, so the business processing request can be preprocessed in advance to facilitate subsequent classification and processing of the business processing request. Specifically, an initial processing request input by the user is obtained, and then the initial processing request is preprocessed. For example, preprocessing may include removing stop words, deleting duplicate information, etc., to obtain preprocessed business processing requests.

S102: Use the text classification model to classify business processing requests and determine the target type of business.

In a possible implementation, the data processing device includes multiple business type databases, and each database pre-collects and stores historical user business problems and processing results corresponding to each business problem. Therefore, when the data processing device obtains the user's business processing request, it can use the text classification model to classify the user's business processing request and determine the target type of business that the user needs to handle, so that the processing results can be obtained from the corresponding type of database.

The text classification model is trained based on the text collection and the business tags corresponding to the text collection, and the trained text classification model is stored in the data processing device in advance. As some examples, the text classification model can be an existing model used for text classification, such as a convolutional neural network, a support vector machine, etc. By encoding the text, extracting features, etc., text classification is finally achieved. The specific implementation principle can be added to the existing text classification model, and will not be introduced here.

For ease of understanding, the training process of the text classification model is introduced below.

First, a training sample set is obtained, where the training sample set includes a text set, and each text in the text set includes a business label, and the business label is used to represent the business type to which the text belongs. The text collection is used as input, and the business labels corresponding to the text collection are used as the expected output of the initial classification model, and the initial classification model is trained to obtain the text classification model.

In specific implementation, the text collection is input into the initial classification model, and the prediction label corresponding to each text in the text collection is obtained. The loss function is determined based on the predicted label and the business label corresponding to the text collection. Among them, the loss function can be used to represent the difference between the prediction label and the business label. When the loss function is larger, it indicates that the difference between the prediction label and the business label is greater, that is, the initial classification model is less accurate. When the loss function is greater than the preset value, the parameters of the initial classification model need to be adjusted, and the text collection is input to the initial classification model after adjusting the parameters, and the subsequent training process is re-executed until the regained loss function is less than or equal to the preset value. value to obtain the trained text classification model.

S103: Perform entity identification on the business processing request and obtain the entity identification result.

Entity recognition refers to identifying entities in a given text and classifying entities, such as time, person name, place name, account number and other types of entities. In this embodiment, the identified entity may be a person's name, bank account number, business type, etc.

In a possible implementation, an entity recognition model based on sequence annotation can be used to perform entity recognition on business processing requests. For example, you can first use common annotation methods such as Begin Inner Other (BIO) to annotate business processing requests with symbolic tokens. Then use convolutional neural networks (CNN) and other models to encode and represent the token sequence, then use a fully connected layer to classify each token in the sequence, and finally use conditional random field (CRF) The model performs label prediction and realizes entity recognition.

It should be noted that the embodiment of the present application does not limit the execution order of step S102 and step S103. That is, step S102 may be executed first and then step S103, or step S103 may be executed first and then step S102, or step S102 and step S102 may be executed simultaneously. S103 does not affect the implementation of the embodiments of this application.

S104: Match the entity recognition results with multiple question templates in the target database to determine the target question.

Multiple business type databases are stored in the data processing equipment, and each database stores multiple problem templates and processing results corresponding to the problem templates. Among them, the problem templates and processing results in the database are obtained by obtaining historical business data. In a possible implementation manner, user processing requests in historical business data can be classified according to business types. For specific classification methods, please refer to the above embodiments, using existing classification models, etc., which will not be described again here. Then, entity recognition, keyword extraction, etc. are performed on processing requests of the same category to obtain question templates corresponding to processing requests of the same category. For the processing results corresponding to the question templates in historical business data, entity recognition, keyword extraction, etc. can also be performed to obtain processed processing results. Thus, a database for each business type can be obtained.

After classifying the business processing requests, you can obtain the target database corresponding to the target type of business, that is, the business of the same type. The target database includes multiple question templates and the processing results corresponding to each question template. The entity recognition results can be matched with the multiple question templates in the target database to determine the target question that matches the entity recognition results.

In a possible implementation, the similarity between the entity recognition result and multiple question templates in the target database can be calculated separately, thereby obtaining multiple similarity indicators. Then the question template corresponding to the maximum value among multiple similarity indicators is determined as the target question. When calculating the similarity between the entity recognition result and the question template, it can be achieved through the minimum edit distance, calculation of cosine similarity, etc.

S105: Send the processing result corresponding to the target problem to the user.

After the target problem is determined, the processing results corresponding to the target problem stored in the target database can be sent to the user, thereby realizing the user's business processing.

Through the method provided by the embodiments of this application, text classification and entity recognition can be used to automatically process the user's business processing request, and obtain the processing result corresponding to the business processing request in the corresponding business database, without the need for staff. Manual processing reduces the time users spend handling business, thereby improving the efficiency of handling business.

Based on the above method embodiments, embodiments of the present application also provide a business processing device based on natural language processing. Referring to Figure 2, Figure 2 is a schematic diagram of a business processing device based on natural language processing provided by an embodiment of the present application.

The device 200 includes:

The obtaining unit 201 is used to obtain the business processing request input by the user;

The classification unit 202 is configured to classify the service processing request using a text classification model and determine the target type of service. The text classification model is trained based on a text collection and a business label corresponding to the text collection;

The identification unit 203 is used to perform entity identification on the business processing request and obtain the entity identification result;

The matching unit 204 is used to match the entity recognition result with multiple question templates in a target database to determine the target question. The target database includes multiple question templates and processing results corresponding to each question template. The target database corresponds to the target type business;

The sending unit 205 is used to send the processing result corresponding to the target question to the user.

In a possible implementation, the training process of the text classification model includes:

Obtain a training sample set, the training sample set includes a text set, and each text in the text set includes a business label;

Taking the text collection as input and the business tags respectively corresponding to the text collection as desired output, an initial classification model is trained to obtain the text classification model.

In a possible implementation, the process of taking the text collection as input, and the business labels corresponding to the text collection as the desired output, and training an initial classification model to obtain the text classification model includes:

Input the text collection into the initial classification model and obtain the prediction label corresponding to the text collection;

Determine the loss function according to the business labels corresponding to the prediction labels and the text collection respectively;

When the loss function is greater than the preset value, adjust the parameters of the initial classification model, and re-execute inputting the text collection into the initial classification model and the subsequent training process until the loss function is less than or equal to the preset value. value to obtain the text classification model.

In a possible implementation, the matching unit 204 is specifically configured to calculate the similarity between the entity recognition result and the multiple question templates respectively to obtain multiple similarity indicators; determine the multiple similarity indexes; The question template corresponding to the maximum value in the similarity index is the target question.

In a possible implementation, the obtaining unit 201 is specifically configured to obtain an initial processing request input by a user; preprocess the initial processing request to obtain the business processing request.

Based on the above method embodiments and device embodiments, embodiments of the present application also provide a business processing device based on natural language processing. The following will be introduced with reference to the accompanying drawings.

Referring to Figure 3, Figure 3 is a schematic diagram of a business processing device based on natural language processing provided by an embodiment of the present application.

The device 300 includes: a memory 301 and a processor 302;

The memory 301 is used to store relevant program codes;

The processor 302 is configured to call the program code to execute the business processing method based on natural language processing described in the above method embodiment.

In addition, embodiments of the present application also provide a computer-readable storage medium, the computer-readable storage medium being used to store a computer program, the computer program being used to perform the business processing based on natural language processing described in the above method embodiments. method.

It should be noted that the business processing methods, devices, equipment and media based on natural language processing provided by this application can be used in the field of artificial intelligence or the field of finance. The above are only examples and do not limit the application fields of the business processing methods, devices, equipment and media based on natural language processing provided in this application.

It should be noted that each embodiment in this specification is described in a progressive manner, and each embodiment focuses on its differences from other embodiments. The same and similar parts between the various embodiments can be referred to each other. In particular, the system or device embodiments are described simply because they are basically similar to the method embodiments. For relevant parts, please refer to the partial description of the method embodiments. The device embodiments described above are only illustrative, in which units or modules illustrated as separate components may or may not be physically separated, and components shown as units or modules may or may not be physical modules, that is, It can be located in one place, or it can also be distributed to multiple network units. Some or all of the units or modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Persons of ordinary skill in the art can understand and implement the method without any creative effort.

It should be understood that in this application, "at least one (item)" refers to one or more, and "plurality" refers to two or more. "And/or" is used to describe the relationship between associated objects, indicating that there can be three relationships. For example, "A and/or B" can mean: only A exists, only B exists, and A and B exist simultaneously. , where A and B can be singular or plural. The character "/" generally indicates that the related objects are in an "or" relationship. “At least one of the following” or similar expressions thereof refers to any combination of these items, including any combination of a single item (items) or a plurality of items (items). For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c" ”, where a, b, c can be single or multiple.

It should also be noted that in this article, relational terms such as first and second are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply that these entities or operations There is no such actual relationship or sequence between them. Furthermore, the terms "comprises," "comprises," or any other variations thereof are intended to cover a non-exclusive inclusion such that a process, method, article, or apparatus that includes a list of elements includes not only those elements, but also those not expressly listed other elements, or elements inherent to the process, method, article or equipment. Without further limitation, an element defined by the statement "comprises a..." does not exclude the presence of additional identical elements in a process, method, article, or apparatus that includes the stated element.

The steps of the methods or algorithms described in conjunction with the embodiments disclosed herein may be implemented directly in hardware, in software modules executed by a processor, or in a combination of both. Software modules may be located in random access memory (RAM), memory, read-only memory (ROM), electrically programmable ROM, electrically erasable programmable ROM, registers, hard disks, removable disks, CD-ROMs, or anywhere in the field of technology. any other known form of storage media.

The above description of the disclosed embodiments enables those skilled in the art to implement or use the present application. Various modifications to these embodiments will be readily apparent to those skilled in the art, and the generic principles defined herein may be practiced in other embodiments without departing from the spirit or scope of the application. Therefore, the present application is not to be limited to the embodiments shown herein but is to be accorded the widest scope consistent with the principles and novel features disclosed herein.

Claims (10)

1. A business processing method based on natural language processing, characterized in that the method includes: Obtain the business processing request input by the user; Use a text classification model to classify the business processing request and determine the target type of business. The text classification model is trained based on a text collection and a business label corresponding to the text collection; Perform entity identification on the business processing request and obtain the entity identification result; Match the entity recognition results with multiple problem templates in a target database to determine the target problem. The target database includes multiple problem templates and processing results corresponding to each problem template. The target database is consistent with the target Corresponding to the type of business; Send the processing results corresponding to the target problem to the user. 2. The method according to claim 1, characterized in that the training process of the text classification model includes: Obtain a training sample set, the training sample set includes a text set, and each text in the text set includes a business label; Taking the text collection as input and the business tags respectively corresponding to the text collection as desired output, an initial classification model is trained to obtain the text classification model. 3. The method according to claim 2, characterized in that the text collection is used as input, and the business tags respectively corresponding to the text collection are used as expected output, and an initial classification model is trained to obtain the text classification. Models, including: Input the text collection into the initial classification model and obtain the prediction label corresponding to the text collection; Determine the loss function according to the business labels corresponding to the prediction labels and the text collection respectively; When the loss function is greater than the preset value, adjust the parameters of the initial classification model, and re-execute inputting the text collection into the initial classification model and the subsequent training process until the loss function is less than or equal to the preset value. value to obtain the text classification model. 4. The method according to claim 1, characterized in that matching the entity recognition result with multiple question templates in the target database to determine the target question includes: Calculate the similarity between the entity recognition result and the multiple question templates respectively to obtain multiple similarity indicators; The question template corresponding to the maximum value among the plurality of similarity indicators is determined to be the target question. 5. The method according to any one of claims 1 to 4, characterized in that said obtaining a business processing request input by a user includes: Get the initial processing request input by the user; Preprocess the initial processing request to obtain the service processing request. 6. A business processing device based on natural language processing, characterized in that the device includes: The acquisition unit is used to obtain the business processing request input by the user; A classification unit, configured to classify the business processing request using a text classification model and determine the target type of business. The text classification model is trained based on a text collection and a business label corresponding to the text collection; An identification unit, used to perform entity identification on the business processing request and obtain the entity identification result; A matching unit, used to match the entity recognition result with multiple question templates in a target database to determine the target question. The target database includes multiple question templates and processing results corresponding to each question template. The target The database corresponds to the target type of business; A sending unit, configured to send the processing results corresponding to the target question to the user. 7. The device according to claim 6, characterized in that the training process of the text classification model includes: Obtain a training sample set, the training sample set includes a text set, and each text in the text set includes a business label; Taking the text collection as input and the business tags respectively corresponding to the text collection as desired output, an initial classification model is trained to obtain the text classification model. 8. The device according to claim 7, wherein the text collection is used as input, and the business tags respectively corresponding to the text collection are used as desired outputs to train an initial classification model to obtain the text classification. The model process includes: Input the text collection into the initial classification model and obtain the prediction label corresponding to the text collection; Determine the loss function according to the business labels corresponding to the prediction labels and the text collection respectively; When the loss function is greater than the preset value, adjust the parameters of the initial classification model, and re-execute inputting the text collection into the initial classification model, and the subsequent training process until the loss function is less than or equal to the preset value. value to obtain the text classification model. 9. A business processing device based on natural language processing, characterized in that the device includes: a memory and a processor; The memory is used to store relevant program codes; The processor is configured to call the program code to execute the business processing method based on natural language processing according to any one of claims 1 to 5. 10. A computer-readable storage medium, characterized in that the computer-readable storage medium is used to store a computer program, and the computer program is used to execute the natural language processing-based method described in any one of claims 1 to 5. Business processing methods.