Product Recommendation Method and Device

Abstract

The present application provides a product recommendation method and apparatus, which can be used in the field of artificial intelligence. The method includes: receiving the income and expenditure data and feature data of a target user; obtaining the net income threshold range, deposit term threshold range and initial adaptation score corresponding to each of a plurality of products to be recommended; determining the intermediate adaptation score of each product to be recommended according to a preset first fixed parameter value, the income and expenditure data, the net income threshold range corresponding to each product to be recommended and the initial adaptation score; determining the target adaptation score of each product to be recommended according to a preset deposit term prediction model, a preset income and expenditure stability evaluation model, the income and expenditure data, the feature data, a preset second fixed parameter value, the intermediate adaptation score corresponding to each product to be recommended and the deposit term threshold range; and sending product recommendation information to the target user according to the target adaptation score of each product to be recommended. The present application can improve the accuracy of product recommendation and thus improve the user experience.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and particularly to a product recommendation method and apparatus. Background Art

[0002] With the rapid development of the financial industry, financial products of banks are constantly being updated, and the number of financial products is increasing. Currently, products are generally recommended and displayed in the order of release time or randomly sorted, etc. Customers need to rely on long-term browsing to make their own judgments when choosing, which is not conducive to customer selection. Summary of the Invention

[0003] In view of the problems in the prior art, this application proposes a product recommendation method and apparatus, which can improve the accuracy of product recommendation and thus improve the user experience.

[0004] To solve the above technical problems, this application provides the following technical solutions:

[0005] In a first aspect, this application provides a product recommendation method, including:

[0006] Receiving the income and expenditure data and characteristic data of a target user;

[0007] Obtaining the net income threshold interval, deposit period threshold interval, and initial adaptation score corresponding to each of multiple products to be recommended;

[0008] Determining the intermediate adaptation score of each product to be recommended according to a preset first fixed parameter value, the income and expenditure data, the net income threshold interval corresponding to each product to be recommended, and the initial adaptation score;

[0009] Determining the target adaptation score of each product to be recommended according to a preset deposit period prediction model, a preset income and expenditure stability evaluation model, the income and expenditure data, characteristic data, a preset second fixed parameter value, the intermediate adaptation score corresponding to each product to be recommended, and the deposit period threshold interval;

[0010] Sending product recommendation information to the target user according to the target adaptation score of each product to be recommended.

[0011] Further, the determining the intermediate adaptation score of each product to be recommended according to a preset first fixed parameter value, the income and expenditure data, the net income threshold interval corresponding to each product to be recommended, and the initial adaptation score includes:

[0012] Determining the net income data of the target user according to the income and expenditure data;

[0013] Determining the target net income threshold interval corresponding to the target user from multiple net income threshold intervals according to the net income data;

[0014] Update the initial adaptation score of the product to be recommended corresponding to the target net income threshold range according to the preset first fixed parameter value to obtain the intermediate adaptation score of the product to be recommended.

[0015] The intermediate adaptation scores of the products to be recommended corresponding to the remaining net income threshold ranges are the initial adaptation scores of the products to be recommended. The remaining net income threshold ranges include: each net income threshold range other than the target net income threshold range among the multiple net income threshold ranges.

[0016] Further, the determining the target adaptation scores of the products to be recommended according to the preset deposit period prediction model, the preset income and expenditure stability evaluation model, the income and expenditure data, the feature data, the preset second fixed parameter value, the intermediate adaptation scores corresponding to the products to be recommended, and the deposit period threshold ranges includes:

[0017] Determine the income and expenditure stability evaluation result of the target user according to the income and expenditure data of the target user and the preset income and expenditure stability evaluation model.

[0018] Determine the occupation demand expectation and age demand expectation of the target user according to the occupation parameter, age parameter of the target user and the preset demand expectation generation rule.

[0019] Determine the predicted deposit period value of the target user according to the income and expenditure stability evaluation result, the occupation demand expectation, the age demand expectation, the gender parameter and the preset deposit period prediction model.

[0020] Determine the target adaptation scores of the products to be recommended according to the predicted deposit period value, the second fixed parameter value, the deposit period threshold ranges corresponding to the products to be recommended respectively, and the intermediate adaptation scores.

[0021] The feature data includes: occupation parameter, age parameter and gender parameter.

[0022] Further, the determining the target adaptation scores of the products to be recommended according to the predicted deposit period value, the second fixed parameter value, the deposit period threshold ranges corresponding to the products to be recommended respectively, and the intermediate adaptation scores includes:

[0023] Determine the target deposit period threshold range corresponding to the target user from the multiple deposit period threshold ranges according to the predicted deposit period value.

[0024] Update the intermediate adaptation score of the product to be recommended corresponding to the target deposit period threshold range according to the second fixed parameter value to obtain the target adaptation score of the product to be recommended.

[0025] The target adaptation scores of the products to be recommended corresponding to the remaining deposit term threshold intervals are the intermediate adaptation scores of the products to be recommended. The remaining deposit term threshold intervals include: each of the deposit term threshold intervals in the multiple deposit term threshold intervals except the target deposit term threshold interval.

[0026] Further, the preset deposit term prediction model is pre-trained by applying a BP model.

[0027] Further, the product recommendation method further includes:

[0028] Obtain a training sample set, which includes: the income and expenditure data of a batch of historical users;

[0029] Apply the training sample set and train based on the density peak clustering algorithm and the MPSI stability evaluation index to obtain the preset income and expenditure stability evaluation model.

[0030] Further, the applying the training sample set and training based on the density peak clustering algorithm and the MPSI stability evaluation index to obtain the preset income and expenditure stability evaluation model includes:

[0031] Train based on the density peak clustering algorithm according to the training sample set to obtain a to-be-verified income and expenditure stability evaluation model;

[0032] Conduct a stability evaluation on the to-be-verified income and expenditure stability evaluation model. If the evaluation passes, determine the to-be-verified income and expenditure stability evaluation model as the income and expenditure stability evaluation model.

[0033] In a second aspect, the present application provides a product recommendation device, including:

[0034] A receiving module, configured to receive the income and expenditure data and feature data of a target user;

[0035] An obtaining module, configured to obtain the net income threshold intervals, deposit term threshold intervals, and initial adaptation scores corresponding to multiple products to be recommended;

[0036] A first determination module, configured to determine the intermediate adaptation scores of each product to be recommended according to a preset first fixed parameter value, the income and expenditure data, the net income threshold intervals and initial adaptation scores corresponding to each product to be recommended;

[0037] A second determination module, configured to determine the target adaptation scores of each product to be recommended according to a preset deposit term prediction model, a preset income and expenditure stability evaluation model, the income and expenditure data, feature data, a preset second fixed parameter value, the intermediate adaptation scores corresponding to each product to be recommended, and the deposit term threshold intervals;

[0038] A push module, configured to send product recommendation information to a target user according to the target adaptation scores of each product to be recommended.

[0039] Further, the first determination module includes:

[0040] A net income determination unit, configured to determine the net income data of the target user according to the revenue and expenditure data;

[0041] A target range determination unit, configured to determine the target net income threshold range corresponding to the target user from multiple net income threshold ranges according to the net income data;

[0042] An update unit, configured to update the initial adaptation score of the product to be recommended corresponding to the target net income threshold range according to a preset first fixed parameter value to obtain the intermediate adaptation score of the product to be recommended;

[0043] The intermediate adaptation scores of the products to be recommended corresponding to the remaining net income threshold ranges are the initial adaptation scores of the products to be recommended, and the remaining net income threshold ranges include: each net income threshold range other than the target net income threshold range among the multiple net income threshold ranges.

[0044] Further, the second determination module includes:

[0045] An evaluation unit, configured to determine the revenue and expenditure stability evaluation result of the target user according to the revenue and expenditure data of the target user and a preset revenue and expenditure stability evaluation model;

[0046] An expectation determination unit, configured to determine the occupational demand expectation and age demand expectation of the target user according to the occupational parameters, age parameters of the target user and a preset demand expectation generation rule;

[0047] A deposit period prediction unit, configured to determine the predicted deposit period value of the target user according to the revenue and expenditure stability evaluation result, occupational demand expectation, age demand expectation, gender parameters and a preset deposit period prediction model;

[0048] An adaptation score determination unit, configured to determine the target adaptation scores of each product to be recommended according to the predicted deposit period value, a second fixed parameter value, the deposit period threshold ranges corresponding to each product to be recommended respectively and the intermediate adaptation scores;

[0049] The feature data includes: occupational parameters, age parameters and gender parameters.

[0050] In a third aspect, the present application provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where the processor implements the product recommendation method when executing the program.

[0051] Fourthly, the present application provides a computer-readable storage medium, on which computer instructions are stored, and when the instructions are executed, the product recommendation method described above is implemented.

[0052] As can be seen from the above technical solutions, the present application provides a product recommendation method and device. Among them, the method includes: receiving the income and expenditure data and characteristic data of a target user; obtaining the net income threshold range, deposit term threshold range and initial adaptation score corresponding to each of a plurality of products to be recommended; determining the intermediate adaptation score of each product to be recommended according to a preset first fixed parameter value, the income and expenditure data, the net income threshold range and the initial adaptation score corresponding to each product to be recommended; determining the target adaptation score of each product to be recommended according to a preset deposit term prediction model, a preset income and expenditure stability evaluation model, the income and expenditure data, the characteristic data, a preset second fixed parameter value, the intermediate adaptation score corresponding to each product to be recommended and the deposit term threshold range; and sending product recommendation information to the target user according to the target adaptation score of each product to be recommended, which can improve the accuracy of product recommendation, and further improve the user experience; specifically, it can display the products to be recommended in combination with information such as the classification of the products to be recommended and user characteristics, can improve the success rate of product recommendation, save the time for the user to browse a large number of deposit products, better understand the user's choices, and at the same time, can also improve the profitability of financial enterprises such as banks. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present application. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0054] Figure 1 is a flowchart of the product recommendation method in the embodiment of the present application;

[0055] Figure 2 is a flowchart of steps 301 to 304 of the product recommendation method in the embodiment of the present application;

[0056] Figure 3 is a flowchart of steps 401 to 404 of the product recommendation method in the embodiment of the present application;

[0057] Figure 4 is a logical diagram of the deposit term prediction model in the application example of the present application;

[0058] Figure 5 is a structural diagram of the product recommendation device in the embodiment of the present application;

[0059] Figure 6 Schematic block diagram of the system composition of the electronic device according to the embodiment of the present application. Specific implementation manners

[0060] In order to enable those skilled in the art to better understand the technical solutions in this specification, the technical solutions in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only a part of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

[0061] Based on this, in order to improve the accuracy of product recommendation and thus improve the user experience, the embodiments of the present application provide a product recommendation device, which may be a server or a client device. The client device may include a smart phone, a tablet electronic device, a network set-top box, a portable computer, a desktop computer, a personal digital assistant (PDA), a vehicle-mounted device, a smart wearable device, etc. Among them, the smart wearable device may include smart glasses, smart watches, smart bracelets, etc.

[0062] In practical applications, the part for product recommendation may be executed on the server side as described above, or all operations may be completed in the client device. Specifically, it can be selected according to the processing capacity of the client device and the limitations of the user usage scenario, etc. The present application does not limit this. If all operations are completed in the client device, the client device may further include a processor.

[0063] The above-mentioned client device may have a communication module (i.e., a communication unit), and may be communicatively connected to a remote server to achieve data transmission with the server. The server may include a server on the task scheduling center side, and may also include a server on an intermediate platform in other implementation scenarios, such as a server on a third-party server platform communicatively linked to the task scheduling center server. The server may include a single computer device, or may include a server cluster composed of multiple servers, or a server structure of a distributed device.

[0064] Any suitable network protocol can be used for communication between the server and the client device, including network protocols that have not been developed as of the filing date of this application. The network protocol can, for example, include TCP / IP protocol, UDP / IP protocol, HTTP protocol, HTTPS protocol, etc. Of course, the network protocol can, for example, also include RPC protocol (Remote Procedure Call Protocol) and REST protocol (Representational State Transfer) used on top of the above protocols, etc.

[0065] It should be noted that the product recommendation method and device disclosed in this application can be used in the financial technology field, and can also be used in any field other than the financial technology field. The application fields of the product recommendation method and device disclosed in this application are not limited.

[0066] Specifically, it will be described through the following various embodiments.

[0067] In order to improve the accuracy of product recommendation and thus improve the user experience, this embodiment provides a product recommendation method whose execution entity is a product recommendation device. The product recommendation device includes but is not limited to a server, as Figure 1 shown, the method specifically includes the following content:

[0068] Step 100: Receive the income and expenditure data and characteristic data of the target user.

[0069] Specifically, all debit card income and expenditure statements within a preset time range under the name of the target user can be obtained as the income and expenditure data of the target user; the preset time range can be set according to actual needs. For example, 1 year or 2 years, etc.

[0070] Step 200: Obtain the net income threshold interval, deposit period threshold interval, and initial adaptation score corresponding to each of the multiple products to be recommended.

[0071] Specifically, the products to be recommended can be deposit products; data analysis and processing can be performed in advance to obtain the combination of the net income threshold interval and the products to be recommended, and the combination of the deposit period threshold interval and the products to be recommended; the initial adaptation score of each product to be recommended can be set in advance according to actual needs; the initial adaptation score S1 of each product to be recommended can be set to be the same in advance; each product to be recommended has its corresponding net income threshold interval, deposit period threshold interval, and initial adaptation score.

[0072] Step 300: Determine the intermediate adaptation score of each product to be recommended according to the preset first fixed parameter value, the income and expenditure data, the net income threshold interval corresponding to each product to be recommended, and the initial adaptation score.

[0073] Specifically, the net income value can be compared with the net income threshold ranges corresponding to each product to be recommended. If it falls within a net income threshold range, the initial adaptation score of the product to be recommended corresponding to this net income threshold range is updated. The large deposit threshold DL, medium deposit threshold ML, and small deposit threshold LL can be preset. The range less than or equal to the small deposit threshold LL, the range greater than the small deposit threshold LL and less than or equal to the medium deposit threshold ML, the range greater than the medium deposit threshold ML and less than or equal to the large deposit threshold DL, and the range greater than the large deposit threshold DL are used as the net income threshold ranges respectively; at least one product to be recommended can correspond to each net income threshold range; the preset first fixed parameter value can be set according to actual needs, and this application does not limit it.

[0074] Step 400: Determine the target adaptation scores of each product to be recommended according to the preset deposit period prediction model, the preset income and expenditure stability evaluation model, the income and expenditure data, the feature data, the preset second fixed parameter value, the intermediate adaptation scores corresponding to each product to be recommended, and the deposit period threshold range.

[0075] Specifically, the preset second fixed parameter value can be set according to actual needs; the first fixed parameter value and the second fixed parameter value can be the same or different; the preset deposit period prediction model can be pre-trained by applying the BP model; the preset income and expenditure stability evaluation model can be pre-trained by applying the density peak clustering algorithm (clustering by fast search and find of density peaks, DPC).

[0076] Step 500: Send product recommendation information to the target user according to the target adaptation scores of each product to be recommended.

[0077] Specifically, each product to be recommended can be sorted from high to low according to the target adaptation score, and the product recommendation information of each product to be recommended is displayed in order.

[0078] To further improve the accuracy of product recommendation, refer to Figure 2 , in an embodiment of this application, step 300 includes:

[0079] Step 301: Determine the net income data of the target user according to the income and expenditure data.

[0080] Step 302: Determine the target net income threshold range corresponding to the target user from multiple net income threshold ranges according to the net income data.

[0081] Step 303: Update the initial adaptation score of the product to be recommended corresponding to the target net income threshold range according to a preset first fixed parameter value to obtain the intermediate adaptation score of the product to be recommended; the intermediate adaptation scores of the products to be recommended corresponding to the remaining net income threshold ranges are the initial adaptation scores of the products to be recommended, and the remaining net income threshold ranges include: each of the net income threshold ranges other than the target net income threshold range among the multiple net income threshold ranges.

[0082] In an example, the first fixed parameter value is 10%. The intermediate adaptation score S can be obtained according to S = S1×(1 + 10%). It can be understood that the initial adaptation scores of the products to be recommended corresponding to the remaining net income threshold ranges remain unchanged and serve as the intermediate adaptation scores.

[0083] To further improve the comprehensiveness of the obtained data and thus improve the accuracy of product recommendation, refer to Figure 3 , in an embodiment of the present application, step 400 includes:

[0084] Step 401: Determine the income and expenditure stability evaluation result of the target user according to the income and expenditure data of the target user and a preset income and expenditure stability evaluation model.

[0085] Specifically, the income and expenditure stability evaluation result is one of: income and expenditure stability, income and expenditure moderation, and income and expenditure instability; the income and expenditure stability evaluation model can output 1 to indicate income and expenditure stability, the income and expenditure stability evaluation model can output 0 to indicate income and expenditure moderation, and the income and expenditure stability evaluation model can output -1 to indicate income and expenditure instability.

[0086] Step 402: Determine the occupational demand expectation and age demand expectation of the target user according to the occupational parameters and age parameters of the target user and a preset demand expectation generation rule.

[0087] Specifically, the preset demand expectation generation rule includes: initializing the occupational demand expectation c = 0, initializing the occupational stability weight array C = [c 1 , c 2 ,..., c i (a weight array established according to the stability degree of different occupations, with different preset weights). Update the occupational demand expectation value of the target user according to the weight corresponding to the occupation of the target user, c = c + 1×c i ; initialize the age demand expectation value d = 0, initialize the demand weight array D = [d 1 , d 2 ,..., d i , where i is equal to 4. First, determine whether it is within the marriage age range (data from census data). If so, update the age demand expectation d = d + 1×d 1, then determine the degree of closeness between his / her age and the national average marriage age (median marriage age). If they are close, continue to update the demand expectation d = d + 1×d 2 . Then determine whether it is within the childbearing age range. If so, update the age demand expectation d = d + 1×d 3 . Then determine the degree of closeness to the average childbearing age. If they are close, continue to update the age demand expectation d = d + 1×d 4 .

[0088] Step 403: Determine the predicted deposit period value of the target user according to the income and expenditure stability evaluation result, occupation demand expectation, age demand expectation, gender parameter and the preset deposit period prediction model

[0089] Specifically, if the target user is male, the gender parameter can be 0; if the target user is female, the gender parameter can be 1. The income and expenditure stability evaluation result, occupation demand expectation, age demand expectation and gender parameter can be input into the preset deposit period prediction model, and the output result of the preset deposit period prediction model is used as the predicted deposit period value of the target user

[0090] Step 404: Determine the target adaptation score of each recommended product according to the predicted deposit period value, the second fixed parameter value, the deposit period threshold interval corresponding to each recommended product, and the intermediate adaptation score; the feature data includes: occupation parameter, age parameter and gender parameter

[0091] In order to further improve the accuracy of determining the target adaptation score and thus improve the accuracy of product recommendation, in an embodiment of the present application, step 404 includes

[0092] Step 4041: Determine the target deposit period threshold interval corresponding to the target user from multiple deposit period threshold intervals according to the predicted deposit period value

[0093] Specifically, the predicted deposit period value can be compared with the net income threshold interval corresponding to each recommended product. If the predicted deposit period value belongs to a net income threshold interval, update the initial adaptation score of the recommended product corresponding to the net income threshold interval. The high deposit period threshold HY, medium deposit period threshold MY and low deposit period threshold LY of the predicted deposit period value and the deposit product can be set in advance. The range less than or equal to the low deposit period threshold LY, the range greater than the low deposit period threshold LY and less than or equal to the medium deposit period threshold MY, the range greater than the medium deposit period threshold MY and less than or equal to the high deposit period threshold HY, and the range greater than the high deposit period threshold HY are respectively used as the net income threshold interval; the target deposit period threshold interval is the deposit period threshold interval where the predicted deposit period value is located

[0094] Step 4042: Update the intermediate adaptation score of the product to be recommended corresponding to the target deposit term threshold range according to the second fixed parameter value to obtain the target adaptation score of the product to be recommended; the target adaptation scores of the products to be recommended corresponding to the remaining deposit term threshold ranges are the intermediate adaptation scores of the products to be recommended, and the remaining deposit term threshold ranges include: each deposit term threshold range other than the target deposit term threshold range among the multiple deposit term threshold ranges.

[0095] In an example, the second fixed parameter value is 10%. The target adaptation score S2 can be obtained according to S2 = S × (1 + 10%). It can be understood that the intermediate adaptation scores of the products to be recommended corresponding to the remaining deposit term threshold ranges remain unchanged and serve as the target adaptation scores.

[0096] To further improve the reliability of the income and expenditure stability evaluation model, and then apply the reliable income and expenditure stability evaluation model to improve the reliability of the income and expenditure stability evaluation results, in an embodiment of the present application, the product recommendation method further includes:

[0097] Step 001: Obtain a training sample set, which includes the income and expenditure data of a batch of historical users.

[0098] Step 002: Apply the training sample set to train based on the density peak clustering algorithm and the MPSI stability evaluation index to obtain the preset income and expenditure stability evaluation model.

[0099] To further improve the reliability of the income and expenditure stability evaluation model, in an embodiment of the present application, step 002 includes:

[0100] Step 0021: Train based on the density peak clustering algorithm according to the training sample set to obtain a to-be-verified income and expenditure stability evaluation model.

[0101] Specifically, the density peak clustering algorithm can be trained according to the training sample set to obtain a to-be-verified income and expenditure stability evaluation model.

[0102] Step 0021: Conduct a stability evaluation on the to-be-verified income and expenditure stability evaluation model. If the evaluation passes, determine the to-be-verified income and expenditure stability evaluation model as the income and expenditure stability evaluation model.

[0103] To further illustrate the present solution, in an application example of the present application, step 002 may include:

[0104] S1: Obtain a training sample set and divide the training sample set into a training sample and a verification sample according to actual needs.

[0105] S2: Initialize the number of training samples TNums, the step size for increasing the number of samples INums, the number of learning times LTimes, and the learning rate LRates.

[0106] S3: Read TNums, read the training samples, calculate the annual net income based on the income and expenditure flow of each user to obtain an array X i＝1,2,...,n =[X 1 ,X 2 ,...,X i , and calculate the sample mean

[0107] S4: Calculate the variance of the annual net income of each user in the training samples to obtain an array s 2 i＝1,2,...,n =[s 2 1 ,s 2 2 ,...,s 2 i , and sort it to obtain an ordered array S i＝1,2,...,n =[S 1 ,S 2 ,...,S n .

[0108] S5: Perform density clustering to form the expected distribution of the training samples. Process the ordered array based on the density peak clustering algorithm. The input of the algorithm is the ordered array of the variance of the annual net income S i＝1,2,...,n and the number of clusters CNums (the number of clusters in this model is 3). The output is the set of user clusters UserCluters = [C 1 ,C 2 ,...,C n , where n is consistent with CNums, and thus the user clusters of different income and expenditure types can be obtained.

[0109] S6: Analyze the distribution of the user clusters and obtain the expected distribution ratio of the training samples. Calculate the number of samples U i(i＝1,2,..,n) in each user cluster, and the total number of samples U = ∑U i(i＝1,2,..,n) . Calculate the distribution ratio of each user cluster to obtain the array E of the expected distribution ratio of the training samples i＝1,2,...,n =[E 1 ,E 2 ,...,E n ,

[0110] S7: Based on the user clusters, perform interval division to obtain the classification intervals. Take the values of the head and tail points in the user clusters, and each user cluster obtains an interval, forming an array of interval objects Record the classification interval formed by the head and tail points of the i-th user cluster.

[0111] S8: Read TNums, read the validation samples, analyze the actual distribution of the validation samples in the classification interval, and obtain the actual distribution ratio of the validation samples. The method is as follows: Calculate and sort to obtain the ordered annual net income variance array S of the validation samples i＝1,2,...,n =[S 1 , S 2 ,..., S n , and based on the interval object array E * i＝1,2,...,n =[E * 1 , E * 2 ,..., E * i cluster S i＝1,2,...,n =[S 1 , S 2 ,..., S n , discard the samples that do not fall into the interval, and count the number of samples U within the interval i(i＝1,2,...,n) , the total number of samples U is the number of validation samples TNums, and calculate the distribution ratio of the validation samples in each interval to obtain the actual distribution ratio array A of the validation samples i＝1,2,...,n =[A 1 , A 2 ,..., A n ,

[0112] Among them, according to the execution process from step S2 to step S8, the parameters required for calculating MPSI will be obtained: E i＝1,2,...,n =[E 1 , E 2 ,..., E n , which is the expected distribution ratio array formed by the training samples in the classification interval, and A i＝1,2,...,n =[A 1 , A 2 ,..., A n , which is the actual distribution ratio array formed by the validation samples in the classification interval.

[0113] S9: Calculate the MPSI index, summarize the consistency between the expected distribution and the actual distribution, and evaluate the model stability: Among them, MPSI is the model stability evaluation index, n refers to the number of distribution intervals, and the value of n in this model is 3. A i(i＝1,...,n) obtained from step S8 is the ratio of the number of validation samples distributed in the i-th interval to the total number of validation samples in the classification interval (obtained in step S7). E i(i＝1,...,n)Obtained from step S6, it is the ratio of the number of training samples distributed in the i-th interval in the classification interval (obtained from step S7) to the total number of training samples.

[0114] S10: Whether MPSI is less than the threshold (temporarily set to 0.25). If so, execute step S11; otherwise, execute S12 to increase the training samples.

[0115] S11: Determine whether the number of learning times is less than the learning rate. If so, execute step S12 to continue training; otherwise, stop training, end, and save the classification interval as the final training result.

[0116] S12: Increase the step size according to the sample quantity, and change the quantities of both the training samples and the validation samples to: TNums = TNums + INums, and then return to execute step S3.

[0117] To further illustrate this solution, this application provides an application example of a product recommendation method, which is specifically described as follows:

[0118] Step S01: Data preprocessing: Deposit products are entered into the product library and classified, and the data is analyzed and processed to obtain a threshold table and combinations of deposit products.

[0119] Step S02: Fund situation assessment; Read the initialized product combinations, and the initial adaptation score for each deposit product is S1. Obtain the income and expenditure statements of all debit cards under the target user within one year, and calculate the net income value I of the target user; Compare the net income value with the large deposit threshold DL, medium deposit threshold ML, and small deposit threshold LL in the threshold table respectively. If it falls into a threshold interval, update the initial adaptation score S1 of the deposit product corresponding to this threshold interval to the intermediate adaptation score S = S1×(1 + 10%); For the deposit products corresponding to the remaining threshold intervals, keep the initial adaptation score unchanged, that is, S = S1.

[0120] Step S03: Deposit term prediction; Based on a deposit term prediction model of a BP neural network with an adaptively adjustable learning rate. The model requires a given training sample set D = {(x 1 , y 1 ), (x 2 , y 2 ),...,(x m , y m ),}, x i ∈R d , y i ∈R l , R d can represent the input data set, R lIt can represent the set of actual deposit years. Let \(l\) represent the deposit year attribute, and \(l\) can be 1. Each input data consists of \(d\) attribute descriptions. It is preset that \(d = 4\), which are the four attributes of the income and expenditure stability assessment value, occupation, age, and gender respectively. The input data in the training samples can be preprocessed by the income and expenditure stability assessment model, occupation parameter preprocessing, age parameter preprocessing, and gender parameter preprocessing, and then input into the deposit year prediction model based on the BP neural network to obtain the deposit year prediction value. As Figure 4 shown, the deposit year prediction model based on the BP neural network includes: an input layer, a hidden layer, and an output layer. The working process of the neurons in the BP neural network is summarized by the following mathematical expressions:

[0121]

[0122] where \(T\) j is the threshold of neuron \(j\), \(w\) ij is the connection weight from neuron \(i\) to \(j\), and \(f\) is the transfer function, and its formula is After iterative training, the BP network can determine the thresholds, connection weights, and learning rate of each layer. After obtaining the real user data, the deposit year prediction model can be applied to predict the user's deposit year, and this value is compared with the high-year threshold \(HY\), medium-year threshold \(MY\), and low-year threshold \(LY\) of the deposit products in the threshold table. If it falls into a threshold interval, the intermediate adaptation score of the deposit product corresponding to this threshold interval is updated to the target adaptation score \(S2 = S×(1 + 10\%)\); for the deposit products corresponding to the remaining threshold intervals, the intermediate adaptation score remains unchanged, that is, \(S2 = S\).

[0123] Specifically, to improve the reliability of the income and expenditure stability assessment model, a part of the samples can be obtained from the training sample set as the training samples and verification samples respectively. Each sample in the training sample set includes: the income and expenditure data of the user. The density peak clustering algorithm can be applied to train based on the training samples. The income and expenditure stability assessment model obtained after training includes three classification intervals and an expected distribution ratio array based on the expected distribution of the training samples in each interval. The actual distribution ratio array can be obtained by applying the actual distribution of the verification samples in the three classification intervals. Then, the model stability assessment index is calculated to summarize the consistency between the expected distribution formed by the training samples and the actual distribution formed by the verification samples, and to evaluate whether the classification intervals trained by the classifier are stable:

[0124]

[0125] where \(n\) represents the number of distribution intervals. In this model, the value of \(n\) is 3, corresponding to the income and expenditure stability, income and expenditure moderation, and income and expenditure instability intervals respectively; \(A\) i(i＝1,...,n)represents the ratio of the number of distribution samples in the i-th classification interval to the total number of verification samples in the actual distribution of verification samples; E i(i＝1,...,n) i(i＝1,...,n) represents the ratio of the number of distribution samples in the i-th classification interval to the total number of training samples in the expected distribution of training samples; this is used to evaluate whether the classification intervals formed by the preset income and expenditure stability evaluation model are stable and reliable. If the MPSI index is greater than the stability threshold (e.g., 0.25), then retrain the income and expenditure stability evaluation model, increase the training quantity by a preset step size, reform the two sample distributions and calculate MPSI again. Only when MPSI is less than 0.25 and the number of learning times is greater than the learning rate, the model is stable, and the classification interval at this time is saved as the final training result; the income and expenditure stability evaluation result is the result output when the model is actually used, while the variance distribution interval is the result obtained from model training; the stable model obtained after training can obtain three variance distribution intervals (classification intervals) that can define users of three types: income and expenditure stability, moderate income and expenditure, and unstable income and expenditure. In actual use, when applying this model, by calculating the variance of its annual net income, it is judged which classification interval it falls into to determine what type of user it is and output the income and expenditure stability evaluation result (classification result).

[0126] Specifically, through the unsupervised learning process of the classifier, the variance distribution interval (classification interval) can be obtained, and then this interval is applied to complete the evaluation of the user's income and expenditure stability in actual use.

[0127] The unsupervised learning process (features) of the classifier lies in combining the relationship between variance characteristics and user income and expenditure data, performing interval division based on the DPC algorithm, and using the MPSI index to judge the stability of the model, achieving an effect of self-learning, independent judgment, and automatic classification. The specific description is as follows: 1) Variance can well reflect the situation of data fluctuations. The stability of users with different income and expenditure types can be quantitatively described by the size of the variance of the annual net income. The smaller the variance, the smaller the fluctuation, indicating that the user's income and expenditure are more stable. Calculate the variance of the annual net income of the income and expenditure data of all users in the sample to obtain an array, and then obtain an ordered array through conventional sorting. The entire array is arranged in order on the number axis. Users with stable income and expenditure have smaller variances, and their points are concentrated on the left side of the number axis, those with moderate stability are concentrated in the middle, and those with instability are concentrated on the right side. 2) Process the ordered array based on the density peak clustering algorithm. The input of the algorithm is the ordered array of the variance of the annual net income and the number of clusters, and the output is the set of user clusters, and user clusters of different income and expenditure types can be obtained. 3) Process the user cluster data. Since the input is already an ordered array in advance, take the nearest and farthest points in the user cluster to generate intervals. This is the process of the initial learning. In order to make the classifier classify more stably and accurately, the model sets the learning rate and calculates the MPSI index. The learning rate is to allow the classifier to reach a sufficient training volume, and by continuously increasing the samples, the contingency of small samples is avoided. At the same time, the MPSI index is used as an auxiliary to summarize the consistency between the previous classification result and the current classification result to judge whether the classifier is stable.

[0128] As can be seen from the above description, the product recommendation method provided by this application example can not only maintain and classify the increasing deposit products, but also display deposit products based on user characteristics, which can improve the accuracy of deposit products, enhance the user experience, and display deposit products according to the adaptation between deposit products and user characteristics, saving the time for users to browse a large number of deposit products.

[0129] At the software level, in order to improve the accuracy of product recommendation and thus enhance the user experience, this application provides an embodiment of a product recommendation device for implementing all or part of the content in the product recommendation method. Refer to Figure 5 The product recommendation device specifically includes the following:

[0130] A receiving module 10, configured to receive the income and expenditure data and feature data of a target user;

[0131] An obtaining module 20, configured to obtain the net income threshold interval, deposit period threshold interval, and initial adaptation score corresponding to each of multiple products to be recommended;

[0132] A first determination module 30, configured to determine the intermediate adaptation score of each product to be recommended according to a preset first fixed parameter value, the income and expenditure data, the net income threshold interval and initial adaptation score corresponding to each product to be recommended;

[0133] A second determination module 40, configured to determine the target adaptation score of each product to be recommended according to a preset deposit period prediction model, a preset income and expenditure stability evaluation model, the income and expenditure data, feature data, a preset second fixed parameter value, the intermediate adaptation score corresponding to each product to be recommended, and the deposit period threshold interval;

[0134] A pushing module 50, configured to send product recommendation information to the target user according to the target adaptation score of each product to be recommended.

[0135] In an embodiment of this application, the first determination module includes:

[0136] A net income determination unit, configured to determine the net income data of the target user according to the income and expenditure data;

[0137] A target interval determination unit, configured to determine the target net income threshold interval corresponding to the target user from multiple net income threshold intervals according to the net income data;

[0138] An updating unit, configured to update the initial adaptation score of the product to be recommended corresponding to the target net income threshold interval according to a preset first fixed parameter value to obtain the intermediate adaptation score of the product to be recommended;

[0139] The intermediate adaptation score of the product to be recommended corresponding to the remaining net income threshold intervals is the initial adaptation score of the product to be recommended, and the remaining net income threshold intervals include: each of the net income threshold intervals other than the target net income threshold interval among the multiple net income threshold intervals.

[0140] In an embodiment of the present application, the second determination module includes:

[0141] An evaluation unit, configured to determine the income and expenditure stability evaluation result of the target user according to the income and expenditure data of the target user and a preset income and expenditure stability evaluation model;

[0142] An expectation determination unit, configured to determine the occupation demand expectation and age demand expectation of the target user according to the occupation parameters, age parameters of the target user and a preset demand expectation generation rule;

[0143] A deposit period prediction unit, configured to determine the predicted value of the deposit period of the target user according to the income and expenditure stability evaluation result, occupation demand expectation, age demand expectation, gender parameter and a preset deposit period prediction model;

[0144] An adaptation score determination unit, configured to determine the target adaptation score of each product to be recommended according to the predicted value of the deposit period, the second fixed parameter value, the deposit period threshold intervals respectively corresponding to each product to be recommended, and the intermediate adaptation score; the feature data includes: occupation parameters, age parameters and gender parameters.

[0145] The embodiments of the product recommendation device provided in this specification can specifically be used to execute the processing flow of the embodiments of the above product recommendation method, and its functions will not be elaborated here. For details, reference can be made to the detailed description of the embodiments of the above product recommendation method.

[0146] As can be seen from the above description, the product recommendation method and device provided in the present application can improve the accuracy of product recommendation, thereby improving the user experience; specifically, it can display the products to be recommended by combining information such as the classification of the products to be recommended and user characteristics, can improve the success rate of product recommendation, save the time for users to browse a large number of deposit products, better understand the choices of users, and at the same time, can also increase the profits of financial enterprises such as banks.

[0147] Figure 6 It is a schematic physical structure diagram of an electronic device provided in an embodiment of the present invention, as Figure 6As shown in the figure, the electronic device may include: a processor 401, a communications interface 402, a memory 403, and a communication bus 404. Among them, the processor 401, the communications interface 402, and the memory 403 communicate with each other through the communication bus 404. The processor 401 may call the logical instructions in the memory 403 to execute the following method: receiving a cloud resource task request; determining, according to the cloud resource task request, a call interface for the cloud resource task, so that a third-party cloud platform corresponding to the call interface executes the cloud resource task corresponding to the cloud resource task request according to the necessary task data.

[0148] In addition, when the logical instructions in the above-mentioned memory 403 are implemented in the form of software functional units and sold or used as independent products, they may be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of the technical solution, may be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in various embodiments of the present invention. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROMs), random access memories (RAMs), magnetic disks, or optical discs that can store program codes.

[0149] This embodiment discloses a computer program product. The computer program product includes a computer program stored on a non-transitory computer-readable storage medium. The computer program includes program instructions. When the program instructions are executed by a computer, the computer can execute the methods provided in the above method embodiments, for example, including: receiving the income and expenditure data and feature data of a target user; obtaining the net income threshold interval, deposit period threshold interval, and initial adaptation score corresponding to each of a plurality of products to be recommended; determining the intermediate adaptation score of each product to be recommended according to a preset first fixed parameter value, the income and expenditure data, the net income threshold interval and the initial adaptation score corresponding to each product to be recommended; determining the target adaptation score of each product to be recommended according to a preset deposit period prediction model, a preset income and expenditure stability evaluation model, the income and expenditure data, the feature data, a preset second fixed parameter value, the intermediate adaptation score corresponding to each product to be recommended, and the deposit period threshold interval; and sending product recommendation information to the target user according to the target adaptation score of each product to be recommended.

[0150] This embodiment provides a computer-readable storage medium storing a computer program that causes a computer to execute the methods provided in the above method embodiments. For example, it includes: receiving the income and expenditure data and feature data of a target user; obtaining the net income threshold range, deposit period threshold range, and initial adaptation score corresponding to each of multiple products to be recommended; determining the intermediate adaptation score of each product to be recommended according to a preset first fixed parameter value, the income and expenditure data, the net income threshold range corresponding to each product to be recommended, and the initial adaptation score; determining the target adaptation score of each product to be recommended according to a preset deposit period prediction model, a preset income and expenditure stability evaluation model, the income and expenditure data, the feature data, a preset second fixed parameter value, the intermediate adaptation score corresponding to each product to be recommended, and the deposit period threshold range; and sending product recommendation information to the target user according to the target adaptation scores of each product to be recommended.

[0151] Those skilled in the art should understand that the embodiments of the present invention can be provided as a method, a system, or a computer program product. Therefore, the present invention can take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0152] The present invention is described with reference to the flowcharts and / or block diagrams of methods, apparatuses (systems), and computer program products according to embodiments of the present invention. It should be understood that each flow and / or block in the flowcharts and / or block diagrams, as well as the combination of flows and / or blocks in the flowcharts and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to the processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing devices to generate a machine, so that the instructions executed by the processor of the computer or other programmable data processing devices generate means for implementing the functions specified in Figure 1 one or more of the flows Figure 1 or multiple flows and / or blocks

[0153] These computer program instructions can also be stored in a computer-readable memory that can direct a computer or other programmable data processing device to work in a specific manner, so that the instructions stored in the computer-readable memory generate a manufactured article including instruction means, and the instruction means implements the functions specified in Figure 1 one or more of the flows Figure 1 or multiple flows and / or blocks

[0154] These computer program instructions can also be loaded onto a computer or other programmable data processing apparatus, so that a series of operation steps are performed on the computer or other programmable apparatus to generate a computer-implemented process, thereby providing instructions for implementing the functions specified in one process or a plurality of processes and / or boxes Figure 1 one process or a plurality of processes and / or boxes Figure 1 steps for the functions specified in one box or a plurality of boxes.

[0155] In the description of the present specification, the description with reference to the terms "one embodiment", "a specific embodiment", "some embodiments", "for example", "example", "specific example", or "some examples", etc. means that the specific features, structures, materials or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In the present specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials or characteristics described may be combined in any one or more embodiments or examples in a suitable manner.

[0156] The above specific embodiments have further elaborated on the objectives, technical solutions, and beneficial effects of the present invention. It should be understood that the above are only specific embodiments of the present invention and are not used to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A product recommendation method, characterized in that, it includes: Receiving the income and expenditure data and characteristic data of the target user; Obtaining the net income threshold interval, deposit period threshold interval and initial adaptation score corresponding to each of multiple products to be recommended; Determining the intermediate adaptation score of each product to be recommended according to the preset first fixed parameter value, the income and expenditure data, the net income threshold interval corresponding to each product to be recommended, and the initial adaptation score; Determining the target adaptation score of each product to be recommended according to the preset deposit period prediction model, the preset income and expenditure stability evaluation model, the income and expenditure data, the characteristic data, the preset second fixed parameter value, the intermediate adaptation score corresponding to each product to be recommended, and the deposit period threshold interval; Sending product recommendation information to the target user according to the target adaptation score of each product to be recommended; Among them, the determining the intermediate adaptation score of each product to be recommended according to the preset first fixed parameter value, the income and expenditure data, the net income threshold interval corresponding to each product to be recommended, and the initial adaptation score includes: Determining the net income data of the target user according to the income and expenditure data; Determining the target net income threshold interval corresponding to the target user from multiple net income threshold intervals according to the net income data; Updating the initial adaptation score of the product to be recommended corresponding to the target net income threshold interval according to the preset first fixed parameter value to obtain the intermediate adaptation score of the product to be recommended; The intermediate adaptation scores of the products to be recommended corresponding to the remaining net income threshold intervals are the initial adaptation scores of the products to be recommended, and the remaining net income threshold intervals include: each net income threshold interval other than the target net income threshold interval among the multiple net income threshold intervals; Among them, the determining the target adaptation score of each product to be recommended according to the preset deposit period prediction model, the preset income and expenditure stability evaluation model, the income and expenditure data, the characteristic data, the preset second fixed parameter value, the intermediate adaptation score corresponding to each product to be recommended, and the deposit period threshold interval includes: Determining the income and expenditure stability evaluation result of the target user according to the income and expenditure data of the target user and the preset income and expenditure stability evaluation model; Determining the occupational demand expectation and age demand expectation of the target user according to the occupational parameter, age parameter of the target user and the preset demand expectation generation rule; Determining the predicted deposit period value of the target user according to the income and expenditure stability evaluation result, the occupational demand expectation, the age demand expectation, the gender parameter and the preset deposit period prediction model; Determining the target adaptation score of each product to be recommended according to the predicted deposit period value, the second fixed parameter value, the deposit period threshold interval corresponding to each product to be recommended, and the intermediate adaptation score; The characteristic data includes: occupational parameter, age parameter and gender parameter.

2. The product recommendation method according to claim 1, characterized in that, the determining the target adaptation score of each product to be recommended according to the predicted deposit period value, the second fixed parameter value, the deposit period threshold interval corresponding to each product to be recommended, and the intermediate adaptation score includes: Determine the target deposit term threshold interval corresponding to the target user from multiple deposit term threshold intervals according to the predicted value of the deposit term; Update the intermediate adaptation score of the product to be recommended corresponding to the target deposit term threshold interval according to the second fixed parameter value to obtain the target adaptation score of the product to be recommended; The target adaptation scores of the products to be recommended corresponding to the remaining deposit term threshold intervals are the intermediate adaptation scores of the products to be recommended, and the remaining deposit term threshold intervals include: each deposit term threshold interval other than the target deposit term threshold interval among the multiple deposit term threshold intervals.

3. The product recommendation method according to claim 1, characterized in that, The preset deposit term prediction model is pre-trained by applying the BP model.

4. The product recommendation method according to claim 1, characterized in that, It further includes: Obtain a training sample set, which includes: the income and expenditure data of a batch of historical users; Apply the training sample set to train based on the density peak clustering algorithm and the MPSI stability evaluation index to obtain the preset income and expenditure stability evaluation model.

5. The product recommendation method according to claim 4, characterized in that, The applying the training sample set to train based on the density peak clustering algorithm and the MPSI stability evaluation index to obtain the preset income and expenditure stability evaluation model includes: Train based on the density peak clustering algorithm according to the training sample set to obtain a to-be-verified income and expenditure stability evaluation model; Perform a stability evaluation on the to-be-verified income and expenditure stability evaluation model. If the evaluation passes, determine the to-be-verified income and expenditure stability evaluation model as the income and expenditure stability evaluation model.

6. A product recommendation device, characterized in that, It includes: A receiving module for receiving the income and expenditure data and feature data of the target user; An obtaining module for obtaining the net income threshold interval, deposit term threshold interval and initial adaptation score corresponding to each of the multiple products to be recommended; A first determination module for determining the intermediate adaptation score of each product to be recommended according to the preset first fixed parameter value, the income and expenditure data, the net income threshold interval and the initial adaptation score corresponding to each product to be recommended; A second determination module for determining the target adaptation score of each product to be recommended according to the preset deposit term prediction model, the preset income and expenditure stability evaluation model, the income and expenditure data, the feature data, the preset second fixed parameter value, the intermediate adaptation score corresponding to each product to be recommended and the deposit term threshold interval; A pushing module for sending product recommendation information to the target user according to the target adaptation score of each product to be recommended; Wherein, the first determination module includes: A net income determination unit for determining the net income data of the target user according to the income and expenditure data; A target interval determination unit for determining the target net income threshold interval corresponding to the target user from multiple net income threshold intervals according to the net income data; An updating unit for updating the initial adaptation score of the product to be recommended corresponding to the target net income threshold interval according to the preset first fixed parameter value to obtain the intermediate adaptation score of the product to be recommended; The intermediate adaptation score of the product to be recommended corresponding to the remaining net income threshold intervals is the initial adaptation score of the product to be recommended. The remaining net income threshold intervals include: each of the net income threshold intervals other than the target net income threshold interval among the multiple net income threshold intervals. Among them, the second determination module includes: An evaluation unit, configured to determine the income and expenditure stability evaluation result of the target user according to the income and expenditure data of the target user and a preset income and expenditure stability evaluation model; An expectation determination unit, configured to determine the occupation demand expectation and age demand expectation of the target user according to the occupation parameter, age parameter of the target user and a preset demand expectation generation rule; A deposit period prediction unit, configured to determine the predicted value of the deposit period of the target user according to the income and expenditure stability evaluation result, occupation demand expectation, age demand expectation, gender parameter and a preset deposit period prediction model; An adaptation score determination unit, configured to determine the target adaptation score of each product to be recommended according to the predicted value of the deposit period, the second fixed parameter value, the deposit period threshold interval corresponding to each product to be recommended, and the intermediate adaptation score; The feature data includes: occupation parameter, age parameter and gender parameter.

7. An electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, wherein, when the processor executes the program, the product recommendation method according to any one of claims 1 to 5 is implemented.

8. A computer-readable storage medium, on which a computer instruction is stored, wherein, when the instruction is executed, the product recommendation method according to any one of claims 1 to 5 is implemented.

9. A computer program product, the computer program product includes a computer program stored on a non-transitory computer-readable storage medium, and the computer program includes program instructions, wherein, when the program instructions are executed by a computer, the product recommendation method according to any one of claims 1 to 5 is implemented.

Images