Service recommendation model optimization method and device, equipment and storage medium thereof

Abstract

The embodiment of the invention belongs to the technical field of big data processing, is applied to a business recommendation scene, and relates to a business recommendation model optimization method and device, equipment and a storage medium thereof. Performing classification processing on the historical business data to obtain representation data corresponding to all users in all feature classification dimensions; identifying the maximum and minimum values of representation data corresponding to all the feature classification dimensions, and obtaining the probability distribution representation of all the feature classification dimensions; and completing adjustment and optimization of the business recommendation model by representing the data extreme value and the probability distribution representation. According to the method, the numerical business data of the individual user on each feature classification dimension can be identified, and the business data numerical representation of the user group can also be represented. The business recommendation model is continuously adjusted and optimized through the numerical representation, so that the business recommendation model is more in line with business user groups.

Description

Technical Field

[0001] The present application relates to the technical field of big data processing, and is applied to business recommendation scenarios, and in particular to a business recommendation model optimization method, device, equipment and storage medium thereof. Background Art

[0002] Large models usually contain more parameters and require a large amount of data for multiple training cycles to achieve good results. Low-rank matrices are used to predict the direction of parameter changes, and one-dimensional vectors are used to predict the amplitude of parameter changes. At the same time, pre-trained weights are frozen, making the scale of training parameters much smaller than directly training the original model. This makes it suitable for use in resource-limited environments.

[0003] However, the method of using small-scale parameters to predict large-scale parameter changes has its own disadvantages. The expressive power of small-scale parameters is limited and cannot accurately describe the changing trend of large-scale parameters. In order to accurately describe the changing trend of large-scale parameters, the existing technology uses low-rank matrices to predict the change in the weight of each feature dimension. However, the process of expanding from low dimensions to high dimensions will result in information loss. It can only calculate the weight change, but cannot accurately obtain the actual feature dimension business data change, resulting in the business recommendation model often not being able to directly combine actual business data in optimization. Summary of the invention

[0004] The purpose of the embodiments of the present application is to propose a business recommendation model optimization method, device, equipment and storage medium thereof, so as to solve the problem that when optimizing the business recommendation model in the prior art, most of the time, the weights of the feature dimensions are optimized, and there is information loss when expanding from low dimensions to high dimensions, and there is often a problem that the adjustment and optimization cannot be directly combined with the actual business data.

[0005] In order to solve the above technical problems, the embodiment of the present application provides a service recommendation model optimization method, which adopts the following technical solutions:

[0006] A method for optimizing a business recommendation model comprises the following steps:

[0007] Step 201, obtaining historical service data of batch users on target services through the target monitoring terminal;

[0008] Step 202: classify the historical business data according to preset feature classification dimensions to obtain corresponding representation data of all users in all feature classification dimensions;

[0009] Step 203, selecting different classification dimensions from all feature classification dimensions in turn as current feature classification dimensions, and identifying whether the representation data of the current feature classification dimension is numerical data;

[0010] Step 204: if the characterization data of the current feature classification dimension is not numerical data, then according to a preset numerical conversion method, the characterization data corresponding to the current feature classification dimension of all users are converted into numerical data, the numerical characterization data corresponding to the current feature classification dimension of all users are obtained, and the maximum value of the characterization data corresponding to the current feature classification dimension is identified, and the probability distribution representation of the current feature classification dimension is obtained by function fitting;

[0011] Step 205: if the characterization data of the current feature classification dimension is numerical data, the maximum value of the characterization data corresponding to the current feature classification dimension is identified, and a probability distribution representation of the current feature classification dimension is obtained by function fitting;

[0012] Step 206, repeatedly executing steps 203 to 205, identifying the maximum values ​​of the characterization data corresponding to all feature classification dimensions respectively, and obtaining the probability distribution representation of all feature classification dimensions;

[0013] Step 207, using the maximum values ​​of the characterization data corresponding to all the feature classification dimensions as business recommendation control parameters, and deploying the probability distribution representation of all the feature classification dimensions as business recommendation reference distribution into a preset business recommendation model to complete the adjustment and optimization of the business recommendation model.

[0014] Furthermore, the target monitoring terminal includes a preset service data recording terminal, and the step of obtaining historical service data of batch users on the target service through the target monitoring terminal specifically includes:

[0015] Based on the user identification information, the historical business data of each user in the batch of users on the target business is obtained from the preset business data recording terminal, wherein the historical business data includes transaction number information, transaction amount information, transaction time information, transaction product information or transaction service information.

[0016] Furthermore, the preset feature classification dimensions include feature classification dimensions based on data attribute fields and feature classification dimensions based on time attribute fields. The steps of classifying the historical business data according to the preset feature classification dimensions to obtain the characterization data corresponding to all users in all feature classification dimensions specifically include:

[0017] According to different data attribute fields, the historical business data is classified and processed to obtain the corresponding characterization data of all users in all feature classification dimensions, wherein the historical business data is classified and processed according to different data attribute fields, including the historical business data is classified and processed in the dimensions of transaction number information, transaction amount information, transaction time information, transaction product information or transaction service information, and,

[0018] The historical business data is classified and processed according to different time attribute fields to obtain the representation data corresponding to all users in all feature classification dimensions, wherein the historical business data is classified and processed according to different time attribute fields, including classifying the historical business data from the dimensions of day, week, month, quarter and year.

[0019] Furthermore, after the step of identifying the maximum values ​​of the characterization data corresponding to all feature classification dimensions, the method further includes:

[0020] Perform statistics on the representation data corresponding to each feature classification dimension to obtain the mode value of the representation data corresponding to each feature classification dimension;

[0021] The proportion of different representation data corresponding to each feature classification dimension is calculated to obtain the proportion value of different representation data corresponding to each feature classification dimension.

[0022] Furthermore, the step of obtaining the probability distribution representation of all feature classification dimensions specifically includes:

[0023] According to the maximum values ​​of the representation data corresponding to all feature classification dimensions, the mode value of the representation data corresponding to each feature classification dimension, and the proportion of different representation data corresponding to each feature classification dimension, the mathematical expectation and variance of the representation data corresponding to all feature classification dimensions are obtained;

[0024] Based on the mathematical expectation and variance, the probability distribution representation of all feature classification dimensions is determined.

[0025] Furthermore, the step of using the maximum values ​​of the characterization data corresponding to all the feature classification dimensions as the service recommendation control parameters specifically includes:

[0026] Obtaining the maximum values ​​of the representation data corresponding to each of the feature classification dimensions;

[0027] The maximum values ​​of the representation data corresponding to each of the feature classification dimensions are set as the recommended upper limit value of the business recommendation model;

[0028] Obtaining the minimum values ​​of the representation data corresponding to all the feature classification dimensions respectively;

[0029] The minimum values ​​of the characterization data corresponding to all the feature classification dimensions are set as the recommended lower limit values ​​of the business recommendation model.

[0030] Furthermore, after executing the step of completing the adjustment and optimization of the business recommendation model, the method further includes:

[0031] Obtaining a service request from a new user for the target service;

[0032] Filtering target service data corresponding to the service request according to the service recommendation control parameter and the service recommendation reference distribution;

[0033] The target service data is fed back to the new user for service recommendation.

[0034] In order to solve the above technical problems, the embodiment of the present application also provides a service recommendation model optimization device, which adopts the following technical solution:

[0035] A business recommendation model optimization device, comprising:

[0036] The historical business data acquisition module is used to obtain the historical business data of batch users on the target business through the target monitoring terminal;

[0037] A classification processing module, used to classify the historical business data according to preset feature classification dimensions, and obtain the corresponding representation data of all users in all feature classification dimensions;

[0038] A numerical data identification module is used to select different classification dimensions from all feature classification dimensions in turn as the current feature classification dimension, and identify whether the representation data of the current feature classification dimension is numerical data;

[0039] The first branch processing module is used for, if the characterization data of the current feature classification dimension is not numerical data, then according to a preset numerical conversion method, performing numerical conversion on the characterization data corresponding to the current feature classification dimension of all users, obtaining the numerical characterization data corresponding to the current feature classification dimension of all users, identifying the maximum value of the characterization data corresponding to the current feature classification dimension, and obtaining the probability distribution representation of the current feature classification dimension by function fitting;

[0040] The second branch processing module is used to identify the maximum value of the characterization data corresponding to the current feature classification dimension if the characterization data of the current feature classification dimension is numerical data, and obtain the probability distribution representation of the current feature classification dimension by function fitting;

[0041] A repeated execution module, used to repeatedly execute steps 203 to 205, identify the maximum values ​​of the characterization data corresponding to all feature classification dimensions, and obtain the probability distribution representation of all feature classification dimensions;

[0042] The adjustment and optimization module is used to use the maximum values ​​of the characterization data corresponding to all the feature classification dimensions as business recommendation control parameters, and deploy the probability distribution representation of all the feature classification dimensions as the business recommendation reference distribution into the preset business recommendation model to complete the adjustment and optimization of the business recommendation model.

[0043] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0044] A computer device includes a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of the above-mentioned business recommendation model optimization method when executing the computer-readable instructions.

[0045] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0046] A computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the business recommendation model optimization method as described above.

[0047] Compared with the prior art, the embodiments of the present application have the following beneficial effects:

[0048] The business recommendation model optimization method described in the embodiment of the present application obtains the historical business data of batch users on the target business; classifies the historical business data to obtain the corresponding characterization data of all users in all feature classification dimensions; identifies the maximum values ​​of the characterization data corresponding to all feature classification dimensions, and obtains the probability distribution representation of all feature classification dimensions; uses the maximum values ​​of the characterization data corresponding to all feature classification dimensions as business recommendation control parameters, and deploys the probability distribution representation of all feature classification dimensions as business recommendation reference distributions to the preset business recommendation model to complete the adjustment and optimization of the business recommendation model. By numerically representing the characterization data of all feature classification dimensions, it is convenient for subsequent analysis and processing, and it is possible to identify the numerical business data of each user in the batch of users on each feature classification dimension, and it is also possible to characterize the numerical representation of the group business data corresponding to the batch of users through the numerical business data of each user in the batch of users on each feature classification dimension. Through the above-mentioned numerical representation, the business recommendation model is continuously adjusted and optimized, so that the business recommendation model is more in line with the business user group. BRIEF DESCRIPTION OF THE DRAWINGS

[0049] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0050] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0051] Figure 2 is a flow chart of an embodiment of a method for optimizing a business recommendation model according to the present application;

[0052] Figure 3 is a structural diagram of an embodiment of a device for optimizing a business recommendation model according to the present application;

[0053] Figure 4 It is a structural diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION

[0054] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0055] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0056] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0057] like Figure 1 As shown, the system architecture 100 may include a terminal device 101, a network 102 and a server 103. The terminal device 101 may be a laptop 1011, a tablet computer 1012 or a mobile phone 1013. The network 102 is used to provide a medium for a communication link between the terminal device 101 and the server 103. The network 102 may include various connection types, such as wired, wireless communication links or optical fiber cables.

[0058] The user can use the terminal device 101 to interact with the server 103 through the network 102 to receive or send messages, etc. Various communication client applications can be installed on the terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0059] The terminal device 101 can be any electronic device with a display screen and supporting web browsing. In addition to a laptop computer 1011, a tablet computer 1012 or a mobile phone 1013, the terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 (Moving Picture Experts Group Audio Layer IV), a laptop computer, a desktop computer, etc.

[0060] The server 103 may be a server that provides various services, such as a background server that provides support for a web page displayed on the terminal device 101 .

[0061] It should be noted that the business recommendation model optimization method provided in the embodiment of the present application is generally executed by a server, and accordingly, the business recommendation model optimization device is generally set in the server.

[0062] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0063] Continue to refer Figure 2 , shows a flow chart of an embodiment of a method for optimizing a business recommendation model according to the present application. The method for optimizing a business recommendation model comprises the following steps:

[0064] Step 201: Obtain historical business data of batch users on target business through the target monitoring terminal.

[0065] In this embodiment, the target monitoring terminal is different according to the target business. For example, if the target business is financial service business such as insurance, loan, lease, etc., the target monitoring terminal can be the financial service business data monitoring terminal of the authoritative regulatory department; if the target business is commodity trading business, the target monitoring terminal can be the commodity trading data recording log terminal built by the shopping platform. The common point is that the target monitoring terminal at least comprehensively records the specific business data of all users in the history when they choose a certain service or purchase a certain commodity. Of course, the business data will also be greatly different according to the different target businesses.

[0066] In fact, the historical business data not only includes numerical data, that is, service prices or product prices, historical service times or historical purchase times. Such business data are undoubtedly numerical data; however, the historical business data may also record service names or product names, service specifications or product specifications. Such business data may be non-numerical data.

[0067] In this embodiment, the historical business data of the batch of users on the target business is obtained through the target monitoring terminal. Specifically, the historical business data of the batch of users on the target business can be obtained from the target monitoring terminal by using a data collection method, for example, by using a crawler tool to collect data from the target monitoring terminal. By obtaining the historical business data of the batch of users on the target business, the business recommendation model can be updated and optimized according to the historical business data of the batch of users, and business recommendations can be made to new users.

[0068] Step 202: classify the historical business data according to preset feature classification dimensions to obtain corresponding representation data of all users in all feature classification dimensions.

[0069] In this embodiment, the historical business data is classified according to the preset feature classification dimensions to obtain the corresponding characterization data of all users in all feature classification dimensions. Specifically, the purpose of step 202 is to classify the historical business data according to all preset feature classification dimensions to obtain the corresponding characterization data of all users in each classification dimension. This facilitates the subsequent adjustment and optimization of the business recommendation model from the perspective of all users, that is, the group.

[0070] Step 203 , select different classification dimensions from all feature classification dimensions in turn as current feature classification dimensions, and identify whether the representation data of the current feature classification dimension is numerical data.

[0071] By identifying whether the data representing each feature classification dimension is numerical data, it is convenient for computers to process and analyze it.

[0072] Step 204, if the characterization data of the current feature classification dimension is not numerical data, then according to a preset numerical conversion method, the characterization data corresponding to all users on the current feature classification dimension are numerically converted to obtain the numerical characterization data corresponding to all users on the current feature classification dimension, and the maximum value of the characterization data corresponding to the current feature classification dimension is identified, and the probability distribution representation of the current feature classification dimension is obtained by function fitting.

[0073] By converting the representation data of the feature classification dimensions into non-numerical data and performing numerical conversion to facilitate computer processing and analysis, the high-dimensional text description is converted into a low-dimensional mathematical representation, and the representation data of all feature classification dimensions are numerically represented to facilitate subsequent analysis and processing.

[0074] In this embodiment, the preset numerical conversion method can be understood. For example, a certain service business includes 5 different service packages, namely Package A, Package B, Package C, Package D, and Package E. Before the numerical conversion, the name fields of the 5 service packages are textual names. The name fields of the 5 service packages are numerically converted and converted into 1, 2, 3, 4 and 5 respectively, that is, the above-mentioned textual names are replaced by numerical representations.

[0075] Step 205: If the characterization data of the current feature classification dimension is numerical data, the maximum value of the characterization data corresponding to the current feature classification dimension is identified, and the probability distribution representation of the current feature classification dimension is obtained by function fitting.

[0076] Step 206, repeatedly executing steps 203 to 205, identifying the maximum values ​​of the characterization data corresponding to all feature classification dimensions respectively, and obtaining the probability distribution representation of all feature classification dimensions.

[0077] By identifying the maximum values ​​of the characterization data corresponding to all feature classification dimensions and obtaining the probability distribution representation of all feature classification dimensions, it is possible to identify the numerical business data of each user in the batch of users on each feature classification dimension, and also to characterize the numerical representation of the group business data corresponding to the batch of users through the numerical business data of each user in the batch of users on each feature classification dimension. This facilitates the computer to identify intuitive individual numerical business data and implicit group numerical business data.

[0078] Step 207, using the maximum values ​​of the characterization data corresponding to all the feature classification dimensions as business recommendation control parameters, and deploying the probability distribution representation of all the feature classification dimensions as business recommendation reference distribution into a preset business recommendation model to complete the adjustment and optimization of the business recommendation model.

[0079] By taking the maximum values ​​of the characterization data corresponding to all the feature classification dimensions as the business recommendation control parameters and deploying the probability distribution representation of all the feature classification dimensions as the business recommendation reference distribution into the preset business recommendation model, the adjustment and optimization of the business recommendation model is completed. With the gradual increase of new users and repeated execution of steps 201 to 207, the business recommendation model is continuously adjusted and optimized, so that the business recommendation model is more in line with the group of business users.

[0080] In this embodiment, the historical business data of batch users on the target business is obtained; the historical business data is classified and processed to obtain the corresponding characterization data of all users in all feature classification dimensions; the maximum values ​​of the characterization data corresponding to all feature classification dimensions are identified, and the probability distribution representation of all feature classification dimensions is obtained; the maximum values ​​of the characterization data corresponding to all feature classification dimensions are used as business recommendation control parameters, and the probability distribution representation of all feature classification dimensions is deployed as a business recommendation reference distribution in the preset business recommendation model to complete the adjustment and optimization of the business recommendation model. By numerically representing the characterization data of all feature classification dimensions, it is convenient for subsequent analysis and processing, and it is possible to identify the numerical business data of each user in the batch of users on each feature classification dimension, and it is also possible to characterize the numerical representation of the group business data corresponding to the batch of users through the numerical business data of each user in the batch of users on each feature classification dimension. Through the above numerical representation, the business recommendation model is continuously adjusted and optimized, so that the business recommendation model is more in line with the business user group.

[0081] In this embodiment, the target monitoring terminal includes a preset business data recording terminal, and the step of obtaining the historical business data of batch users on the target business through the target monitoring terminal specifically includes: based on user identification information, obtaining the historical business data of each user in the batch of users on the target business from the preset business data recording terminal, wherein the historical business data includes transaction number information, transaction amount information, transaction time information, transaction product information or transaction service information.

[0082] By acquiring the historical service data of batch users on the target service, the service recommendation model is updated and optimized according to the historical service data of the batch users, and services are recommended to new users.

[0083] In this embodiment, the preset feature classification dimensions include feature classification dimensions based on data attribute fields and feature classification dimensions based on time attribute fields. The step of classifying the historical business data according to the preset feature classification dimensions to obtain the representation data corresponding to all users in all feature classification dimensions specifically includes: classifying the historical business data according to different data attribute fields to obtain the representation data corresponding to all users in all feature classification dimensions, wherein the classification of the historical business data according to different data attribute fields includes classifying the historical business data in the dimensions of transaction number information, transaction amount information, transaction time information, transaction product information or transaction service information, and classifying the historical business data in the dimensions of time attribute fields to obtain the representation data corresponding to all users in all feature classification dimensions, wherein the classification of the historical business data according to the dimensions of time attribute fields includes classifying the historical business data in the dimensions of day, week, month, quarter and year.

[0084] According to all preset feature classification dimensions, the historical business data is classified and processed to obtain the representation data of all users corresponding to each classification dimension, so as to facilitate the subsequent adjustment and optimization of the business recommendation model from the perspective of multiple feature dimensions and groups.

[0085] In this embodiment, after the step of identifying the maximum values ​​of the representation data corresponding to all feature classification dimensions, the method also includes: performing statistics on the representation data corresponding to each feature classification dimension to obtain the mode value of the representation data corresponding to each feature classification dimension; performing proportion calculations on the different representation data corresponding to each feature classification dimension to obtain the proportion values ​​of the different representation data corresponding to each feature classification dimension.

[0086] In this embodiment, the step of obtaining the probability distribution representation of all feature classification dimensions specifically includes: obtaining the mathematical expectation and variance of the representation data corresponding to all feature classification dimensions according to the maximum value of the representation data corresponding to all feature classification dimensions, the mode value of the representation data corresponding to each feature classification dimension, and the proportion of different representation data corresponding to each feature classification dimension; based on the mathematical expectation and variance, determining the probability distribution representation of all feature classification dimensions.

[0087] Through mathematical comprehensive analysis, the maximum value of the characterization data, the mode value of the characterization data, and the proportion of different characterization data corresponding to each feature classification dimension are obtained, and then the mathematical expectation and variance are further obtained to determine the probability distribution representation of all feature classification dimensions. This is to optimize and guide the business recommendation model based on the probability distribution representation.

[0088] In this embodiment, the step of using the maximum values ​​of the representation data corresponding to all the feature classification dimensions as the business recommendation control parameters specifically includes: obtaining the maximum values ​​of the representation data corresponding to all the feature classification dimensions; setting the maximum values ​​of the representation data corresponding to all the feature classification dimensions as the recommended upper limit value of the business recommendation model; obtaining the minimum values ​​of the representation data corresponding to all the feature classification dimensions; setting the minimum values ​​of the representation data corresponding to all the feature classification dimensions as the recommended lower limit value of the business recommendation model.

[0089] By obtaining the maximum and minimum values ​​of the characterization data corresponding to all the feature classification dimensions, they are set as the recommended upper limit and recommended lower limit of the business recommendation model. This ensures that when the business recommendation model recommends business data on the corresponding feature dimension, it does not exceed the recommended upper limit and recommended lower limit, and only recommends within the interval formed by the recommended upper limit and recommended lower limit.

[0090] In this embodiment, after executing the step of completing the adjustment and optimization of the business recommendation model, the method also includes: obtaining the new user's business request for the target business; filtering out the target business data corresponding to the business request based on the business recommendation control parameters and the business recommendation reference distribution; and feeding back the target business data to the new user for business recommendation.

[0091] In this embodiment, the method further includes: according to a preset update optimization strategy, repeatedly executing steps 201 to 207, re-obtaining the probability distribution representation of all feature classification dimensions, and optimizing and updating the business recommendation model based on the probability distribution representation.

[0092] Specifically, the preset update optimization strategy includes repeating steps 201 to 207 when the number of new users reaches a certain threshold, or when the last update time reaches a preset time interval threshold from the current time, thereby achieving continuous optimization and adjustment of the business recommendation model.

[0093] This application obtains the historical business data of batch users on the target business; classifies the historical business data to obtain the corresponding characterization data of all users in all feature classification dimensions; identifies the maximum values ​​of the characterization data corresponding to all feature classification dimensions, and obtains the probability distribution representation of all feature classification dimensions; uses the maximum values ​​of the characterization data corresponding to all feature classification dimensions as business recommendation control parameters, and deploys the probability distribution representation of all feature classification dimensions as business recommendation reference distribution to the preset business recommendation model to complete the adjustment and optimization of the business recommendation model. By numerically representing the characterization data of all feature classification dimensions, it is convenient for subsequent analysis and processing, and can not only identify the numerical business data of each user in the batch of users on each feature classification dimension, but also can characterize the numerical representation of the group business data corresponding to the batch of users through the numerical business data of each user in the batch of users on each feature classification dimension. Through the above numerical representation, the business recommendation model is continuously adjusted and optimized, so that the business recommendation model is more in line with the business user group.

[0094] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Among them, artificial intelligence (AI) is the theory, method, technology and application system that uses digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0095] AI basic technologies generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing technology, operation / interaction systems, mechatronics, etc. AI software technologies mainly include computer vision technology, robotics technology, biometrics technology, speech processing technology, natural language processing technology, and machine learning / deep learning.

[0096] In the embodiment of the present application, by obtaining the historical business data of batch users on the target business; classifying the historical business data, obtaining the corresponding characterization data of all users in all feature classification dimensions; identifying the maximum values ​​of the characterization data corresponding to all feature classification dimensions, and obtaining the probability distribution representation of all feature classification dimensions; using the maximum values ​​of the characterization data corresponding to all feature classification dimensions as business recommendation control parameters, and deploying the probability distribution representation of all feature classification dimensions as business recommendation reference distribution to the preset business recommendation model, the adjustment and optimization of the business recommendation model is completed. By numerically representing the characterization data of all feature classification dimensions, it is convenient for subsequent analysis and processing, and it is possible to identify the numerical business data of each user in the batch of users in each feature classification dimension, and it is also possible to characterize the numerical representation of the group business data corresponding to the batch of users through the numerical business data of each user in the batch of users in each feature classification dimension. Through the above numerical representation, the business recommendation model is continuously adjusted and optimized, so that the business recommendation model is more in line with the business user group.

[0097] Further references Figure 3 , as a response to the above Figure 2 The present application provides an embodiment of a device for optimizing a business recommendation model. Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0098] like Figure 3 As shown, the business recommendation model optimization device 300 described in this embodiment includes: a historical business data acquisition module 301, a classification processing module 302, a numerical data identification module 303, a first branch processing module 304, a second branch processing module 305, a repeated execution module 306 and an adjustment optimization module 307. Among them:

[0099] The historical service data acquisition module 301 is used to acquire the historical service data of batch users on the target service through the target monitoring terminal;

[0100] The classification processing module 302 is used to classify the historical business data according to the preset feature classification dimensions, and obtain the corresponding representation data of all users in all feature classification dimensions;

[0101] A numerical data identification module 303 is used to select different classification dimensions from all feature classification dimensions in turn as the current feature classification dimension, and identify whether the representation data of the current feature classification dimension is numerical data;

[0102] The first branch processing module 304 is used to convert the corresponding characterization data of all users on the current feature classification dimension into numerical data according to a preset numerical conversion method if the characterization data of the current feature classification dimension is not numerical data, obtain the numerical characterization data corresponding to all users on the current feature classification dimension, identify the maximum value of the characterization data corresponding to the current feature classification dimension, and obtain the probability distribution representation of the current feature classification dimension by function fitting;

[0103] The second branch processing module 305 is used to identify the maximum value of the characterization data corresponding to the current feature classification dimension if the characterization data of the current feature classification dimension is numerical data, and obtain the probability distribution representation of the current feature classification dimension by function fitting;

[0104] Repeating execution module 306, used to repeatedly execute step 203 to step 205, identify the maximum values ​​of the characterization data corresponding to all feature classification dimensions, and obtain the probability distribution representation of all feature classification dimensions;

[0105] The adjustment and optimization module 307 is used to use the maximum values ​​of the characterization data corresponding to all the feature classification dimensions as business recommendation control parameters, and deploy the probability distribution representation of all the feature classification dimensions as the business recommendation reference distribution into the preset business recommendation model to complete the adjustment and optimization of the business recommendation model.

[0106] This application obtains the historical business data of batch users on the target business; classifies the historical business data to obtain the corresponding characterization data of all users in all feature classification dimensions; identifies the maximum values ​​of the characterization data corresponding to all feature classification dimensions, and obtains the probability distribution representation of all feature classification dimensions; uses the maximum values ​​of the characterization data corresponding to all feature classification dimensions as business recommendation control parameters, and deploys the probability distribution representation of all feature classification dimensions as business recommendation reference distribution to the preset business recommendation model to complete the adjustment and optimization of the business recommendation model. By numerically representing the characterization data of all feature classification dimensions, it is convenient for subsequent analysis and processing, and can not only identify the numerical business data of each user in the batch of users on each feature classification dimension, but also can characterize the numerical representation of the group business data corresponding to the batch of users through the numerical business data of each user in the batch of users on each feature classification dimension. Through the above numerical representation, the business recommendation model is continuously adjusted and optimized, so that the business recommendation model is more in line with the business user group.

[0107] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0108] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0109] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0110] The computer device 4 includes a memory 4a, a processor 4b, and a network interface 4c which are interconnected through a system bus. It should be noted that: Figure 4 Only a computer device 4 having a component memory 4a, a processor 4b, and a network interface 4c is shown, but it should be understood that it is not required to implement all the components shown, and more or fewer components may be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to a microprocessor, an application specific integrated circuit (ASIC), a field programmable gate array (FPGA), a digital processor (DSP), an embedded device, etc.

[0111] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0112] The memory 4a includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 4a can be an internal storage unit of the computer device 4, such as a hard disk or memory of the computer device 4. In other embodiments, the memory 4a can also be an external storage device of the computer device 4, such as a plug-in hard disk equipped on the computer device 4, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. Of course, the memory 4a can also include both the internal storage unit of the computer device 4 and its external storage device. In this embodiment, the memory 4a is generally used to store the operating system and various application software installed on the computer device 4, such as a computer-readable instruction of a business recommendation model optimization method, etc. In addition, the memory 4a can also be used to temporarily store various types of data that have been output or are to be output.

[0113] The processor 4b may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip in some embodiments. The processor 4b is generally used to control the overall operation of the computer device 4. In this embodiment, the processor 4b is used to run computer-readable instructions stored in the memory 4a or process data, such as computer-readable instructions for running the business recommendation model optimization method.

[0114] The network interface 4c may include a wireless network interface or a wired network interface. The network interface 4c is generally used to establish a communication connection between the computer device 4 and other electronic devices.

[0115] The computer device proposed in this embodiment belongs to the field of big data processing technology and is applied to business recommendation scenarios. This application obtains the historical business data of batch users on the target business; classifies the historical business data to obtain the corresponding characterization data of all users in all feature classification dimensions; identifies the maximum values ​​of the characterization data corresponding to all feature classification dimensions, and obtains the probability distribution representation of all feature classification dimensions; uses the maximum values ​​of the characterization data corresponding to all feature classification dimensions as business recommendation control parameters, and deploys the probability distribution representation of all feature classification dimensions as business recommendation reference distributions to the preset business recommendation model to complete the adjustment and optimization of the business recommendation model. By numerically representing the characterization data of all feature classification dimensions, it is convenient for subsequent analysis and processing, and can identify the numerical business data of each user in the batch of users on each feature classification dimension, and can also characterize the numerical representation of the group business data corresponding to the batch of users through the numerical business data of each user in the batch of users on each feature classification dimension. Through the above-mentioned numerical representation, the business recommendation model is continuously adjusted and optimized, so that the business recommendation model is more in line with the business user group.

[0116] The present application also provides another implementation, namely, providing a computer-readable storage medium, wherein the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by a processor to enable the processor to perform the steps of the business recommendation model optimization method as described above.

[0117] The computer-readable storage medium proposed in this embodiment belongs to the field of big data processing technology and is applied to business recommendation scenarios. This application obtains the historical business data of batch users on the target business; classifies the historical business data to obtain the corresponding characterization data of all users in all feature classification dimensions; identifies the maximum values ​​of the characterization data corresponding to all feature classification dimensions, and obtains the probability distribution representation of all feature classification dimensions; uses the maximum values ​​of the characterization data corresponding to all feature classification dimensions as business recommendation control parameters, and deploys the probability distribution representation of all feature classification dimensions as business recommendation reference distributions to the preset business recommendation model to complete the adjustment and optimization of the business recommendation model. By numerically representing the characterization data of all feature classification dimensions, it is convenient for subsequent analysis and processing, and can identify the numerical business data of each user in the batch of users on each feature classification dimension, and can also characterize the numerical representation of the group business data corresponding to the batch of users through the numerical business data of each user in the batch of users on each feature classification dimension. Through the above-mentioned numerical representation, the business recommendation model is continuously adjusted and optimized, so that the business recommendation model is more in line with the business user group.

[0118] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0119] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A business recommendation model optimization method, characterized in that: The steps include: Step 201, obtaining historical service data of batch users on target services through the target monitoring terminal; Step 202: classify the historical business data according to preset feature classification dimensions to obtain corresponding representation data of all users in all feature classification dimensions; Step 203, selecting different classification dimensions from all feature classification dimensions in turn as current feature classification dimensions, and identifying whether the representation data of the current feature classification dimension is numerical data; Step 204: if the characterization data of the current feature classification dimension is not numerical data, then according to a preset numerical conversion method, the characterization data corresponding to the current feature classification dimension of all users are converted into numerical data, the numerical characterization data corresponding to the current feature classification dimension of all users are obtained, and the maximum value of the characterization data corresponding to the current feature classification dimension is identified, and the probability distribution representation of the current feature classification dimension is obtained by function fitting; Step 205: if the characterization data of the current feature classification dimension is numerical data, the maximum value of the characterization data corresponding to the current feature classification dimension is identified, and a probability distribution representation of the current feature classification dimension is obtained by function fitting; Step 206, repeatedly executing steps 203 to 205, identifying the maximum values ​​of the characterization data corresponding to all feature classification dimensions respectively, and obtaining the probability distribution representation of all feature classification dimensions; Step 207, using the maximum values ​​of the characterization data corresponding to all the feature classification dimensions as business recommendation control parameters, and deploying the probability distribution representation of all the feature classification dimensions as business recommendation reference distribution into a preset business recommendation model to complete the adjustment and optimization of the business recommendation model.

2. The business recommendation model optimization method according to claim 1, characterized in that: The target monitoring terminal includes a preset business data recording terminal, and the step of obtaining historical business data of batch users on the target business through the target monitoring terminal specifically includes: Based on the user identification information, the historical business data of each user in the batch of users on the target business is obtained from the preset business data recording terminal, wherein the historical business data includes transaction number information, transaction amount information, transaction time information, transaction product information or transaction service information.

3. The business recommendation model optimization method according to claim 1, characterized in that: The preset feature classification dimensions include feature classification dimensions based on data attribute fields and feature classification dimensions based on time attribute fields. The step of classifying the historical business data according to the preset feature classification dimensions to obtain the characterization data corresponding to all users in all feature classification dimensions specifically includes: According to different data attribute fields, the historical business data is classified and processed to obtain the corresponding characterization data of all users in all feature classification dimensions, wherein the historical business data is classified and processed according to different data attribute fields, including the historical business data is classified and processed in the dimensions of transaction number information, transaction amount information, transaction time information, transaction product information or transaction service information, and, The historical business data is classified and processed according to different time attribute fields to obtain the representation data corresponding to all users in all feature classification dimensions, wherein the historical business data is classified and processed according to different time attribute fields, including classifying the historical business data from the dimensions of day, week, month, quarter and year.

4. The business recommendation model optimization method according to claim 1, characterized in that: After the step of identifying the maximum values ​​of the characterization data corresponding to all feature classification dimensions, the method further includes: Perform statistics on the representation data corresponding to each feature classification dimension to obtain the mode value of the representation data corresponding to each feature classification dimension; The proportion of different representation data corresponding to each feature classification dimension is calculated to obtain the proportion value of different representation data corresponding to each feature classification dimension.

5. The business recommendation model optimization method according to claim 4, characterized in that: The step of obtaining the probability distribution representation of all feature classification dimensions specifically includes: According to the maximum values ​​of the representation data corresponding to all feature classification dimensions, the mode value of the representation data corresponding to each feature classification dimension, and the proportion of different representation data corresponding to each feature classification dimension, the mathematical expectation and variance of the representation data corresponding to all feature classification dimensions are obtained; Based on the mathematical expectation and variance, the probability distribution representation of all feature classification dimensions is determined.

6. The business recommendation model optimization method according to claim 1 or 5, characterized in that: The step of using the maximum values ​​of the characterization data corresponding to all the feature classification dimensions as the business recommendation control parameters specifically includes: Obtaining the maximum values ​​of the representation data corresponding to each of the feature classification dimensions; The maximum values ​​of the representation data corresponding to each of the feature classification dimensions are set as the recommended upper limit value of the business recommendation model; Obtaining the minimum values ​​of the representation data corresponding to all the feature classification dimensions respectively; The minimum values ​​of the characterization data corresponding to all the feature classification dimensions are set as the recommended lower limit values ​​of the business recommendation model.

7. The business recommendation model optimization method according to claim 1, characterized in that: After executing the step of completing the adjustment and optimization of the business recommendation model, the method further includes: Obtaining a service request from a new user for the target service; Filtering target service data corresponding to the service request according to the service recommendation control parameter and the service recommendation reference distribution; The target service data is fed back to the new user for service recommendation.

8. A business recommendation model optimization device, characterized in that: include: The historical business data acquisition module is used to obtain the historical business data of batch users on the target business through the target monitoring terminal; A classification processing module, used to classify the historical business data according to preset feature classification dimensions, and obtain the corresponding representation data of all users in all feature classification dimensions; A numerical data identification module is used to select different classification dimensions from all feature classification dimensions in turn as the current feature classification dimension, and identify whether the representation data of the current feature classification dimension is numerical data; The first branch processing module is used for, if the characterization data of the current feature classification dimension is not numerical data, then according to a preset numerical conversion method, performing numerical conversion on the characterization data corresponding to the current feature classification dimension of all users, obtaining the numerical characterization data corresponding to the current feature classification dimension of all users, identifying the maximum value of the characterization data corresponding to the current feature classification dimension, and obtaining the probability distribution representation of the current feature classification dimension by function fitting; The second branch processing module is used to identify the maximum value of the characterization data corresponding to the current feature classification dimension if the characterization data of the current feature classification dimension is numerical data, and obtain the probability distribution representation of the current feature classification dimension by function fitting; A repeated execution module, used to repeatedly execute steps 203 to 205, identify the maximum values ​​of the characterization data corresponding to all feature classification dimensions, and obtain the probability distribution representation of all feature classification dimensions; The adjustment and optimization module is used to use the maximum values ​​of the characterization data corresponding to all the feature classification dimensions as business recommendation control parameters, and to deploy the probability distribution representation of all the feature classification dimensions as the business recommendation reference distribution into the preset business recommendation model to complete the adjustment and optimization of the business recommendation model.

9. A computer device, characterized in that: It comprises a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the business recommendation model optimization method as described in any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, and when the computer-readable instructions are executed by the processor, the steps of the business recommendation model optimization method according to any one of claims 1 to 7 are implemented.

Images