Personal financial product recommendation method and device

Abstract

This application discloses a method and device for recommending personal financial products, relating to the financial field. The method comprises: performing a cluster analysis on a target customer and multiple non-target customers based on their basic information and financial product purchase data to obtain a customer grouping containing the target customer; obtaining multiple financial products purchased by each non-target customer in the customer grouping and their corresponding sales data; performing a cluster analysis on the multiple financial products based on the target customer's purchase propensity data and the sales data to obtain multiple financial product groups; and recommending financial products to the target customer based on the purchase quantity of each financial product contained in each financial product grouping. The application utilizes a large customer sample base, which helps improve the accuracy of financial product recommendation results. Replacing subjective analysis by sales personnel with computer analysis helps improve the objectivity of financial product recommendation results.

Description

Technical Field

[0001] The present application relates to the financial field, and specifically to a method and device for recommending personal financial products. Background Art

[0002] In recent years, commercial banks' personal wealth management products have become increasingly popular among consumers, leading to increasingly fierce competition among banks. As a strategic resource for commercial banks, they need to develop personal wealth management products that appeal to customers and encourage them to purchase and hold them long-term.

[0003] Commercial banks typically use research reports to analyze customer preferences for personal financial products. First, they collect customer reports, either online or offline. Then, they use the reports to conduct a basic analysis of customer personal information and their understanding of financial products, understanding their gender, age, education level, monthly income, and understanding of financial products. Finally, based on customer personality, habits, and psychological characteristics, and using customer market segmentation and statistical analysis, they identify different customer types and analyze the characteristics of the personal financial products preferred by each type of customer.

[0004] However, the method of understanding customers by filling out research reports has the following disadvantages: first, the sample of customers surveyed is small and cannot represent the preferences of the majority of customers; second, the analysis of the research report is subjective, and the professional levels of different business personnel are different, so the results of the analysis will vary to a certain extent. Summary of the Invention

[0005] In order to solve the problems existing in the prior art, in a first aspect, the present application provides a method for recommending personal financial products, comprising:

[0006] Performing cluster analysis on the target customer and multiple non-target customers based on their basic information data and financial product purchase data to obtain the customer group to which the target customer belongs;

[0007] Obtaining multiple financial products purchased by each non-target customer in the customer group and their corresponding sales data;

[0008] performing cluster analysis on the plurality of financial products based on the target customer's purchase tendency data and the sales data to obtain a plurality of financial product groups;

[0009] Financial products are recommended to the target customers based on the purchase quantity of each financial product included in each financial product group.

[0010] In one embodiment, the cluster analysis is performed on the target customer and a plurality of non-target customers based on their basic information data and financial product purchase data to obtain the customer group to which the target customer belongs, including:

[0011] Obtaining basic information data and financial product purchase data of the target customer and multiple non-target customers;

[0012] Filtering the financial product purchase data based on a preset first screening rule to obtain valid financial product purchase data;

[0013] Build customer characteristics for each customer based on their corresponding basic information data and effective financial product purchase data;

[0014] Perform cluster analysis on the customers based on the customer characteristics to obtain the customer group to which the target customers belong.

[0015] In one embodiment, after constructing the customer characteristics of each customer based on the basic information data and valid financial product purchase data corresponding to each customer, the method further includes:

[0016] Screening the customer features based on filtering and recursive feature elimination;

[0017] Based on the principal component analysis method, feature extraction and feature mapping are performed on each customer feature to obtain the mapped customer features.

[0018] In one embodiment, obtaining the plurality of financial products purchased by each non-target customer in the customer group and the corresponding sales data includes:

[0019] Obtaining multiple financial products purchased by each non-target customer in the customer group and their corresponding original sales data;

[0020] The original sales data corresponding to each of the financial products is screened based on a preset second screening rule to obtain valid sales data corresponding to each of the financial products.

[0021] In one embodiment, the cluster analysis of the plurality of financial products based on the target customer's purchase tendency data and the sales data to obtain a plurality of financial product groups includes:

[0022] Determining the product characteristics of each of the financial products based on the target customer's purchasing tendency data and the effective sales data;

[0023] Cluster analysis is performed on the multiple financial products based on the product characteristics to obtain multiple financial product groups.

[0024] In one embodiment, determining the product features of each of the financial products based on the target customer's purchasing tendency data and the sales data includes:

[0025] Obtaining the target customer's purchasing tendency data, wherein the purchasing tendency data includes long-term holding and trying new products;

[0026] The product characteristics of each financial product are respectively constructed according to the feature construction rules corresponding to the purchase tendency data and the effective sales data corresponding to each financial product.

[0027] In one embodiment, the recommending financial products to the target customer based on the purchase quantity of each financial product included in each financial product group includes:

[0028] Obtaining the purchase quantity of each financial product included in each financial product group respectively, and determining the total purchase quantity of each financial product group based on the purchase quantity;

[0029] Each financial product in the financial product group with the largest total purchase quantity is recommended to the target customer.

[0030] In a second aspect, the present application provides a personal financial product recommendation device, comprising:

[0031] A customer grouping module is used to perform cluster analysis on customers based on the basic information data and financial product purchase data of target customers and multiple non-target customers to obtain the customer group to which the target customers belong;

[0032] A financial product sales data acquisition module is used to acquire multiple financial products purchased by each non-target customer in the customer group and their corresponding sales data;

[0033] a financial product grouping module, configured to perform cluster analysis on the plurality of financial products based on the target customer's purchase tendency data and the sales data to obtain a plurality of financial product groups;

[0034] The financial product recommendation module is used to recommend financial products to the target customers based on the purchase quantity of each financial product included in each financial product group.

[0035] In one embodiment, the customer grouping module includes:

[0036] A customer information acquisition unit, configured to acquire basic information data and financial product purchase data of the target customer and a plurality of non-target customers;

[0037] a purchase data screening unit, configured to screen the financial product purchase data based on a preset first screening rule to obtain valid financial product purchase data;

[0038] A customer feature construction unit is used to construct customer features for each customer based on their corresponding basic information data and valid financial product purchase data;

[0039] The customer grouping unit is configured to perform cluster analysis on the customers based on the customer characteristics to obtain the customer group to which the target customer belongs.

[0040] In one embodiment, the customer grouping module further includes a customer feature processing unit configured to:

[0041] Screening the customer features based on filtering and recursive feature elimination;

[0042] Based on the principal component analysis method, feature extraction and feature mapping are performed on each customer feature to obtain the mapped customer features.

[0043] In one embodiment, the financial product sales data acquisition module includes:

[0044] a sales data acquisition unit, configured to acquire a plurality of financial products purchased by each non-target customer in the customer group and the corresponding original sales data;

[0045] The sales data filtering unit is used to filter the original sales data corresponding to each of the financial products based on a preset second filtering rule to obtain valid sales data corresponding to each of the financial products.

[0046] In one embodiment, the financial product grouping module includes:

[0047] a financial product feature construction unit, configured to determine the product features of each of the financial products based on the target customer's purchase tendency data and the effective sales data;

[0048] The financial product grouping unit is used to perform cluster analysis on the multiple financial products based on the product characteristics to obtain multiple financial product groups.

[0049] In one embodiment, the financial product feature construction unit is specifically configured to:

[0050] Obtaining the target customer's purchasing tendency data, wherein the purchasing tendency data includes long-term holding and trying new products;

[0051] The product characteristics of each financial product are respectively constructed according to the feature construction rules corresponding to the purchase tendency data and the effective sales data corresponding to each financial product.

[0052] In one embodiment, the financial product recommendation module includes:

[0053] a purchase quantity statistics unit, configured to respectively obtain the purchase quantity of each financial product included in each financial product group, and determine the total purchase quantity of each financial product group based on the purchase quantity;

[0054] The financial product recommendation unit is used to recommend each financial product in the financial product group with the largest total purchase quantity to the target customer.

[0055] In a third aspect, the present application provides a computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements any one of the personal financial product recommendation methods provided in the present application.

[0056] In a fourth aspect, the present application provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, it implements any personal financial product recommendation method provided in the present application.

[0057] The personal financial product recommendation method and device of the present application uses self-learning modeling and classification technology to classify customers based on their purchase records of personal financial products, thereby obtaining customer groups related to the target customer's financial product purchasing behavior. The device then obtains the purchase records of financial products belonging to the target customer's customer group, constructs different product features based on the target customer's current purchase tendency, and uses self-learning modeling and classification technology to obtain multiple financial product groups. Finally, the device recommends personal financial products to the target customer in the corresponding group based on the customer's willingness to purchase. This application overcomes the shortcomings of existing solutions that use research reports to assess customer willingness to purchase personal financial products. Due to the large customer sample base, it helps to improve the accuracy of financial product recommendation results. By replacing subjective analysis by sales personnel with computer analysis, the device helps to improve the objectivity of financial product recommendation results. BRIEF DESCRIPTION OF THE DRAWINGS

[0058] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for the embodiments or the description of the prior art. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative work. In the drawings:

[0059] Figure 1 A schematic diagram of the personal financial product recommendation method provided in this application.

[0060] Figure 2 Another schematic diagram of the personal financial product recommendation method provided for this application.

[0061] Figure 3Another schematic diagram of the personal financial product recommendation method provided for this application.

[0062] Figure 4 Another schematic diagram of the personal financial product recommendation method provided for this application.

[0063] Figure 5 Another schematic diagram of the personal financial product recommendation method provided for this application.

[0064] Figure 6 Another schematic diagram of the personal financial product recommendation method provided for this application.

[0065] Figure 7 A schematic diagram of the personal financial product recommendation device provided in this application.

[0066] Figure 8 This is another schematic diagram of the personal financial product recommendation device provided in this application.

[0067] Figure 9 This is another schematic diagram of the personal financial product recommendation device provided in this application.

[0068] Figure 10 This is another schematic diagram of the personal financial product recommendation device provided in this application.

[0069] Figure 11 This is another schematic diagram of the personal financial product recommendation device provided in this application.

[0070] Figure 12 This is another schematic diagram of the personal financial product recommendation device provided in this application.

[0071] Figure 13 A schematic diagram of a computer device provided in this application. DETAILED DESCRIPTION

[0072] To make the purpose, technical solutions and advantages of the embodiments of the present invention more clear, the embodiments of the present invention are further described in detail below with reference to the accompanying drawings. Here, the exemplary embodiments of the present invention and their descriptions are used to explain the present invention, but are not intended to limit the present invention.

[0073] It should be noted that the personal financial product recommendation method and device of the present application can be used in the financial field, and can also be used in any field other than the financial field. The present application does not limit the application field of the personal financial product recommendation method and device.

[0074] The acquisition, storage, use, and processing of data in the technical solution of this application comply with the relevant provisions of national laws and regulations.

[0075] The user information in the embodiments of this application is obtained through legal and compliant channels, and the acquisition, storage, use, and processing of the user information are authorized and agreed upon by the customer.

[0076] In the first aspect, the present application provides a method for recommending personal financial products, such as Figure 1 As shown, the method includes:

[0077] Step S101 : performing cluster analysis on the target customer and a plurality of non-target customers based on their basic information data and financial product purchase data to obtain a customer group to which the target customer belongs.

[0078] Target and non-target customers are both customers targeted by the company and will be collectively referred to as customers in subsequent instructions. The greater the number of target and non-target customers, the better the financial product recommendations. Basic information data includes but is not limited to customer number, customer gender, customer level, customer age, etc.; financial product purchase data includes but is not limited to financial product number, customer number, customer purchase time, and customer purchase amount.

[0079] This step uses cluster analysis based on a large amount of customer basic information and wealth management product data to identify similarities between different customers. This cluster analysis yields multiple customer groups. Customers within the same group share similar wealth management product purchasing behaviors, such as similarities in investment amount range, risk tolerance, and preferred wealth management product types.

[0080] Step S102: Acquire multiple financial products purchased by each non-target customer in the customer group and their corresponding sales data.

[0081] Specifically, multiple customer groups are obtained in step S101. The financial management styles of customers in the same customer group are similar. Therefore, this step directly obtains multiple financial management products purchased by customers in the customer group where the target customer is located and their corresponding sales data, so as to obtain multiple financial management products that the customer group tends to purchase based on these data.

[0082] This step only obtains the financial products purchased by non-target customers in the target customer's group and their corresponding sales data, and does not consider the financial products purchased by the target customer himself and their corresponding sales data.

[0083] Step S103 , performing cluster analysis on the plurality of financial products based on the target customer's purchase tendency data and the sales data to obtain a plurality of financial product groups.

[0084] This step constructs product features corresponding to each financial product based on the multiple financial products and their corresponding sales data obtained in step S102, and then performs cluster analysis on each financial product based on the product features to obtain multiple financial product groups.

[0085] This step takes into account the target customer's purchasing tendency. The purchasing tendency is selected by the customer. Different purchasing tendencies correspond to different methods of constructing product features, which will be explained in detail in subsequent embodiments.

[0086] Step S104: recommending financial products to the target customers based on the purchase quantity of each financial product included in each financial product group.

[0087] This step counts the purchase quantity of each financial product group determined in step S1032. The greater the purchase quantity, the more likely it is that it meets the customer's purchase intention. This step recommends the financial products in the financial product group with the largest purchase quantity to the target customer.

[0088] In one embodiment, if Figure 2 As shown, step S101 is to perform cluster analysis on the target customer and multiple non-target customers based on their basic information data and financial product purchase data to obtain the customer group to which the target customer belongs, including:

[0089] Step S1011: Obtain basic information data and financial product purchase data of the target customer and multiple non-target customers.

[0090] Step S1012: Filter the financial product purchase data based on a preset first filtering rule to obtain valid financial product purchase data.

[0091] Specifically, the financial product purchase data includes but is not limited to the financial product number, customer number, customer purchase time, customer purchase amount, etc. This step filters out missing, abnormal, or meaningless financial product purchase data based on the preset first screening rules. The first screening rules include but are not limited to:

[0092] (1) Eliminate the purchase data of financial products whose product numbers do not exist in the basic information of financial products stored in the bank;

[0093] (2) Eliminate the purchase data of financial products whose purchase time is earlier than the launch time of the financial products;

[0094] (3) Eliminate the purchase data of financial products whose purchase time is later than the launch time of the financial products.

[0095] After eliminating missing, abnormal and meaningless financial product purchase data according to the first screening rule, valid financial product purchase data is obtained.

[0096] Step S1013: construct customer characteristics for each customer based on the basic information data and valid financial product purchase data corresponding to each customer.

[0097] This step requires building customer profiles for both target and non-target customers. The following uses target customers as an example. Building customer profiles for other customers can be found in the implementation.

[0098] Specifically, assuming that the target customer has purchased financial products for n years, then it is necessary to use the customer data with experience in purchasing financial products greater than or equal to n years and less than or equal to n+m years to construct feature data (n and m are both positive integers, and m is a preset positive integer). The constructed customer features specifically include: customer gender, age, customer level, total number of financial products purchased by the customer, total amount of financial products purchased by the customer, average rate of return of financial products purchased by the customer, average number of years the customer has purchased financial products, the proportion of high-risk and high-return financial products held by the customer, and how many years the customer has purchased financial products; number of financial products purchased by the customer in year T, amount of financial products purchased by the customer in year T, proportion of high-risk and high-return financial products purchased by the customer in year T, average rate of return of financial products purchased by the customer in year T, highest rate of return of financial products purchased by the customer in year T, lowest rate of return of financial products purchased by the customer in year T; customer The number of wealth management products purchased in year T-1, the amount of wealth management products purchased by the customer in year T-1, the proportion of high-risk, high-return wealth management products purchased by the customer in year T-1, the average rate of return on wealth management products purchased by the customer in year T-1, the highest rate of return on wealth management products purchased by the customer in year T-1, the lowest rate of return on wealth management products purchased by the customer in year T-1, etc.; the number of wealth management products purchased by the customer in year Tn, the amount of wealth management products purchased by the customer in year Tn, the proportion of high-risk, high-return wealth management products purchased by the customer in year Tn, the average rate of return on wealth management products purchased by the customer in year Tn, the highest rate of return on wealth management products purchased by the customer in year Tn, the lowest rate of return on wealth management products purchased by the customer in year Tn. The number of features constructed for each customer is 6n+15.

[0099] The above customer characteristics are constructed mainly based on mathematical statistics methods to analyze and process the basic information data of customers and the effective financial product purchase data to obtain the customer characteristics of each customer.

[0100] Step S1014: performing cluster analysis on the customers based on the customer characteristics to obtain the customer group to which the target customer belongs.

[0101] The K-means algorithm is used to perform unsupervised cluster analysis on each customer within a customer group based on their characteristics. Initially, customers are divided into N clusters. The iteration termination criteria are: the cluster center remains unchanged, and the cluster assigned to each data point remains unchanged. After clustering is complete, N customer groups are obtained. The group in which the target customer is located is the group consisting of customers related to the target customer's financial product purchasing behavior.

[0102] In one embodiment, if Figure 3 As shown, in step S1013, after constructing the customer characteristics of each customer based on the basic information data and valid financial product purchase data corresponding to each customer, the following is also included:

[0103] Step S1015: screening the customer features based on filtering method and recursive feature elimination method.

[0104] Specifically, feature selection is performed on the constructed customer features, using methods including: removing customer features with the smallest value changes and single customer features with the lowest importance evaluation through filtering, and removing customer features with poor performance in the feature subset through recursive feature elimination.

[0105] Step S1016 , performing feature extraction and feature mapping on each customer feature based on principal component analysis to obtain mapped customer features.

[0106] Specifically, principal component analysis (PCA) is used to extract features from the customer features screened in step S1015, and a new feature space is mapped to reduce the dimension of the customer features and improve computational efficiency.

[0107] In one embodiment, if Figure 4 As shown, step S102, obtaining multiple financial products purchased by each non-target customer in the customer group and their corresponding sales data, includes:

[0108] Step S1021, obtaining multiple financial products purchased by each non-target customer in the customer group and their corresponding original sales data.

[0109] The original sales data corresponding to the wealth management products here include customer-related wealth management product purchase data, namely the wealth management product number, customer number, customer purchase time, customer purchase amount, etc., as well as information such as the bank-related wealth management product online time and wealth management product offline time.

[0110] Step S101 and related embodiments detail the process of deriving multiple customer groups based on the basic information and financial product purchase data of a target customer and multiple non-target customers. The customers in the customer group to which the target customer belongs share similar financial product purchasing behavior. Therefore, this step obtains the financial products purchased by each non-target customer in the customer group to which the target customer belongs, as well as the raw sales data corresponding to these financial products, to further analyze and identify financial products that the target customer is likely to purchase.

[0111] It should be noted that, to avoid interference, this step only obtains information on non-target customers in the customer group to which the target customer belongs, and does not include the multiple financial products purchased by the target customer himself and their corresponding original sales data.

[0112] Step S1022: Filter the original sales data corresponding to each of the financial products based on a preset second filtering rule to obtain valid sales data corresponding to each of the financial products.

[0113] Specifically, this step filters the original sales data corresponding to each of the wealth management products obtained in step S1021 based on the second screening rule, and removes the original sales data corresponding to each of the wealth management products that are missing, abnormal, and meaningless, to obtain valid sales data. The second screening rule includes but is not limited to:

[0114] (1) Eliminate the purchase data of financial products whose numbers do not exist in the basic information of financial products stored in the bank;

[0115] (2) Eliminate the purchase data of financial products whose purchase time is earlier than the launch time of the financial products;

[0116] (3) Eliminate the purchase data of financial products whose purchase time is later than the launch time of the financial products.

[0117] In one embodiment, if Figure 5 As shown, in step S103, cluster analysis is performed on the multiple financial products based on the purchase tendency data of the target customers and the sales data to obtain multiple financial product groups, including:

[0118] Step S1031 : determining the product characteristics of each of the financial products according to the target customer's purchase tendency data and the effective sales data.

[0119] Since the purchase tendency data is different, the construction method of the product features of the financial product is different. Therefore, this step needs to determine the construction method of the product features of the financial product based on the purchase tendency data of the target customer. Specifically, step S1031 can be implemented by the following steps:

[0120] Step 1: Obtain the target customer's purchasing tendency data, including long-term holding and trying new products. The purchasing tendency data is selected by the target customer himself.

[0121] Step 2: constructing product features of each financial product based on the feature construction rules corresponding to the purchase tendency data and the effective sales data corresponding to each financial product.

[0122] (1) The purchase tendency data indicates long-term holding, meaning that customers tend to purchase long-term financial products. In this case, customer loyalty should be considered. Based on the effective sales data of each financial product obtained in step S102, product features are constructed for each financial product. The following description will use a specific financial product as an example; the construction of product features for other financial products can be similarly implemented.

[0123] Specifically, assuming that the maximum number of times a customer purchases a wealth management product is m, through the feature construction method, the main features of each wealth management product are: the average length of time a customer holds the wealth management product / the length of time the wealth management product has been online, the average number of purchases by customers who have purchased the wealth management product, the number of customers who repeatedly purchase the wealth management product / the number of customers who have purchased the wealth management product; the number of customers who have purchased the wealth management product once in total / the number of customers who have purchased the wealth management product, the number of customers who have purchased the wealth management product twice in total / the number of customers who have purchased the wealth management product, the number of customers who have purchased the wealth management product m times in total / the number of customers who have purchased the wealth management product; the number of customers who have held the wealth management product for more than 30% of the length of time the wealth management product has been online / the number of customers who have purchased the wealth management product, the number of customers who have held the wealth management product for more than 50% of the length of time the wealth management product has been online / the number of customers who have purchased the wealth management product, the number of customers who have held the wealth management product for more than 70% of the length of time the wealth management product has been online / the number of customers who have purchased the wealth management product 5+4m features are generated for each wealth management product. The first and second holding periods are: average holding period (m-1th - average holding period); average holding period (m-1th - average holding period); average holding period (m-2th - average holding period); average holding period (m-1th - average holding period); average balance (m-2th - average balance); average balance (m-1th - average balance); average balance (m-2th - average balance); average balance (m-2th - average balance); average balance (m-3th - average balance); average balance (m-1th - average balance). Generate 5+4m features for each wealth management product.

[0124] (2) Purchase propensity data refers to trying new products, i.e., customers tend to purchase newer financial products that they have not purchased before or have purchased less frequently. In this case, customer preference should be considered. Based on the effective sales data of each financial product obtained in step S102, product features are constructed for each financial product. The following description will use a specific financial product as an example; the construction of product features for other financial products can be similarly implemented.

[0125] Specifically, assuming that the maximum online time of a wealth management product is n years, and the maximum number of times a customer purchases the wealth management product is m, the main features of each wealth management product obtained through the feature construction method are: the average number of purchases by customers who have purchased the wealth management product; the number of first purchases by customers in the first month after the wealth management product is launched, the number of first purchases by customers in the second month after the wealth management product is launched, ..., the number of first purchases by customers in the 12th month after the wealth management product is launched; the number of first purchases of the wealth management product in the first year, the number of first purchases of the wealth management product in the second year, ..., the number of first purchases of the wealth management product in the nth year; the number of first purchases of the wealth management product in the second year / the number of first purchases of the wealth management product in the first year, the number of first purchases of the wealth management product in the third year / the number of first purchases of the wealth management product in the second year, ..., the number of first purchases of the wealth management product in the nth year The number of first purchases of the wealth management product in a year / the number of first purchases of the wealth management product in the n-1th year; the average amount of a customer's first purchase, the average amount of a customer's second purchase, ..., the average amount of a customer's m-th purchase; the average amount of a customer's second purchase / the average amount of a customer's first purchase, the average amount of a customer's third purchase / the average amount of a customer's second purchase, ..., the average amount of a customer's m-th purchase / the average amount of a customer's m-1-th purchase; the number of customers who made the first purchase, the number of customers who made the second purchase, ..., the number of customers who made the m-th purchase; the average time between a customer's second purchase and their first purchase, the average time between a customer's third purchase and their second purchase, ..., the average time between a customer's m-th purchase and their m-1-th purchase. For each wealth management product, 10+2n+4m features are generated.

[0126] The above product characteristics are mainly constructed based on the analysis and processing of the effective sales data of financial products using mathematical statistics methods to obtain the product characteristics of each financial product.

[0127] Step S1032: performing cluster analysis on the plurality of financial products based on the product characteristics to obtain a plurality of financial product groups.

[0128] Specifically, based on the product features obtained in step S1031, an unsupervised cluster analysis of the financial products is performed using the K-means algorithm. Initially, the financial products are divided into M categories. The iteration termination conditions are: the cluster center does not change, and the cluster assigned to each data point remains unchanged. After clustering is complete, M financial product groups are obtained.

[0129] In actual applications, manual confirmation can also be used to assist, with the customer manager confirming the classification results in step S1032, screening out incorrectly classified financial products, and sending the product features of these financial products to step S1013 for reclassification until the classification is correct or no classification is possible.

[0130] In one embodiment, if Figure 6 As shown, step S104, recommending financial products to the target customer based on the purchase quantity of each financial product included in each financial product group, includes:

[0131] Step S1041 , respectively obtaining the purchase quantity of each financial product included in each financial product group, and determining the total purchase quantity of each financial product group based on the purchase quantity.

[0132] Specifically, for each wealth management product group, the purchase quantity of each wealth management product included in the group is counted. The purchase quantity of a wealth management product in this step includes the total amount of the wealth management product purchased by all customers between the time the wealth management product goes online and the time it goes offline. For example, this can be measured by a combination of purchase amount and holding period. For a wealth management product group, the sum of the purchase quantities of all wealth management products included in the group is the total purchase quantity of the wealth management product group.

[0133] Step S1042: recommend each financial product in the financial product group with the largest total purchase quantity to the target customer.

[0134] The above embodiment obtains M financial product groups based on the cluster analysis algorithm. This step sorts these M financial product groups based on the total purchase quantity, and recommends the financial product group with the largest total purchase quantity to the target customer as the financial product group with the greatest customer purchase intention.

[0135] This embodiment categorizes wealth management products based on customer purchase intent, helping commercial banks select and recommend the most popular wealth management products to customers. In addition to the total purchase quantity mentioned above, in actual applications, parameters such as average purchase quantity can also be used to assess customer purchase intent, thereby improving the accuracy of wealth management product recommendations.

[0136] The personal financial product recommendation method of the present application uses self-learning modeling and classification technology to classify customers based on their purchase records of personal financial products, thereby obtaining customer groups related to the target customer's financial product purchasing behavior. The method then obtains the purchase records of the financial products in the customer group to which the target customer belongs, constructs different product features based on the target customer's current purchase tendency, and uses self-learning modeling and classification technology to obtain multiple financial product groups. Finally, the method recommends personal financial products in the corresponding group to the target customer based on the customer's willingness to purchase. This method overcomes the shortcomings of existing solutions that use research reports to assess customer willingness to purchase personal financial products. Due to the large customer sample base, it helps to improve the accuracy of financial product recommendation results. By replacing subjective analysis by sales personnel with computer analysis, it helps to improve the objectivity of financial product recommendation results.

[0137] Based on the same inventive concept, the embodiments of the present application also provide a personal financial product recommendation device, which can be used to implement the method described in the above embodiments, as described in the following embodiments. Since the principle of solving the problem by the personal financial product recommendation device is similar to that of the personal financial product recommendation method, the implementation of the personal financial product recommendation device can refer to the implementation of the personal financial product recommendation method, and the repeated parts will not be repeated. As used below, the term "unit" or "module" can be a combination of software and / or hardware that implements the predetermined function. Although the system described in the following embodiments is preferably implemented in software, implementation in hardware, or a combination of software and hardware, is also possible and conceived.

[0138] In the second aspect, the present application provides a personal financial product recommendation device, such as Figure 7 As shown, the device includes:

[0139] A customer grouping module 201 is configured to perform cluster analysis on a target customer and a plurality of non-target customers based on their basic information data and financial product purchase data, and obtain a customer group to which the target customer belongs;

[0140] The financial product sales data acquisition module 202 is used to acquire multiple financial products purchased by each non-target customer in the customer group and their corresponding sales data;

[0141] The financial product grouping module 203 is configured to perform cluster analysis on the plurality of financial products based on the target customer's purchase tendency data and the sales data to obtain a plurality of financial product groups;

[0142] The financial product recommendation module 204 is configured to recommend financial products to the target customer based on the purchase quantity of each financial product included in each financial product group.

[0143] In one embodiment, if Figure 8As shown, the customer grouping module 201 includes:

[0144] The customer information acquisition unit 2011 is used to acquire basic information data and financial product purchase data of the target customer and multiple non-target customers;

[0145] a purchase data screening unit 2012, configured to screen the financial product purchase data based on a preset first screening rule to obtain valid financial product purchase data;

[0146] The customer feature construction unit 2013 is used to construct customer features for each customer based on the basic information data and valid financial product purchase data corresponding to each customer;

[0147] The customer grouping unit 2014 is configured to perform cluster analysis on the customers based on the customer characteristics to obtain the customer group to which the target customer belongs.

[0148] In one embodiment, if Figure 9 As shown, the customer grouping module 201 further includes a customer feature processing unit 2015, which is used to:

[0149] Screening the customer features based on filtering and recursive feature elimination;

[0150] Based on the principal component analysis method, feature extraction and feature mapping are performed on each customer feature to obtain the mapped customer features.

[0151] In one embodiment, if Figure 10 As shown, the financial product sales data acquisition module 202 includes:

[0152] The sales data acquisition unit 221 is used to acquire multiple financial products purchased by each non-target customer in the customer group and their corresponding original sales data;

[0153] The sales data filtering unit 2022 is configured to filter the original sales data corresponding to each of the financial products based on a preset second filtering rule to obtain valid sales data corresponding to each of the financial products.

[0154] In one embodiment, if Figure 11 As shown, the financial product grouping module 203 includes:

[0155] The financial product feature construction unit 2031 is configured to determine the product features of each of the financial products based on the target customer's purchase tendency data and the effective sales data;

[0156] The financial product grouping unit 2032 is configured to perform cluster analysis on the plurality of financial products based on the product features to obtain a plurality of financial product groups.

[0157] In one embodiment, the financial product feature construction unit 2031 is specifically configured to:

[0158] Obtaining the target customer's purchasing tendency data, wherein the purchasing tendency data includes long-term holding and trying new products;

[0159] The product characteristics of each financial product are respectively constructed according to the feature construction rules corresponding to the purchase tendency data and the effective sales data corresponding to each financial product.

[0160] In one embodiment, if Figure 12 As shown, the financial product recommendation module 204 includes:

[0161] The purchase quantity statistics unit 2041 is used to obtain the purchase quantity of each financial product included in each financial product group, and determine the total purchase quantity of each financial product group based on the purchase quantity;

[0162] The financial product recommendation unit 2042 is configured to recommend each financial product in the financial product group with the largest total purchase quantity to the target customer.

[0163] The personal financial product recommendation device of the present application uses self-learning modeling and classification technology to classify customers based on their purchase records of personal financial products, thereby obtaining customer groups related to the target customer's financial product purchasing behavior. The device then obtains the purchase records of financial products belonging to the target customer's customer group, constructs different product features based on the target customer's current purchase propensity, and uses self-learning modeling and classification technology to obtain multiple financial product groups. Finally, the device recommends personal financial products to the target customer in the corresponding group based on the customer's willingness to purchase. This application overcomes the shortcomings of existing solutions that use research reports to assess customer willingness to purchase personal financial products. Due to the large customer sample base, the device helps improve the accuracy of financial product recommendation results. Furthermore, computer analysis replaces subjective analysis by sales personnel, thereby improving the objectivity of financial product recommendation results.

[0164] In a third aspect, the present application also provides a computer device, see Figure 13 , the electronic device 100 specifically includes:

[0165] A central processing unit (CPU) 110 , a memory (memory) 120 , a communication module (Communications) 130 , an input unit 140 , an output unit 150 and a power supply 160 .

[0166] The memory 120, communication module 130, input unit 140, output unit 150, and power supply 160 are respectively connected to the central processing unit 110. The memory 120 stores a computer program, which the central processing unit 110 can call. When the central processing unit 110 executes the computer program, all steps of the personal financial product recommendation method in the above embodiment are implemented.

[0167] In one embodiment, the present application further provides a computer-readable storage medium for storing a computer program, wherein the computer program is executable by a processor. When the computer program is executed by the processor, any of the personal financial product recommendation methods provided by the present invention is implemented.

[0168] In one embodiment, an embodiment of the present invention further provides a computer program product, which includes a computer program. When the computer program is executed by a processor, it implements any of the personal financial product recommendation methods provided in the above embodiments.

[0169] It will be understood by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Thus, the present invention may take the form of an entirely hardware embodiment, an entirely software embodiment, or an embodiment combining software and hardware. Furthermore, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to magnetic disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0170] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowcharts and / or block diagrams, as well as combinations of processes 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 a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowcharts and / or block diagrams. Figure 1 a process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0171] These computer program instructions may 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 produce an article of manufacture comprising an instruction device, which implements the process Figure 1a process or multiple processes and / or boxes Figure 1 The function specified in one or more boxes.

[0172] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operational steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing the instructions executed on the computer or other programmable device for implementing the process. Figure 1 a process or multiple processes and / or boxes Figure 1 A step that specifies a function in one or more boxes.

[0173] The specific embodiments described above further illustrate the objectives, technical solutions and beneficial effects of the present invention in detail. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A personal financial product recommendation method, characterized in that: include: Performing cluster analysis on the target customer and multiple non-target customers based on their basic information data and financial product purchase data to obtain the customer group to which the target customer belongs; Obtaining multiple financial products purchased by each non-target customer in the customer group and their corresponding sales data; performing cluster analysis on the plurality of financial products based on the target customer's purchase tendency data and the sales data to obtain a plurality of financial product groups; recommending financial products to the target customers based on the purchase quantity of each financial product included in each financial product group; The obtaining of the plurality of financial products purchased by each non-target customer in the customer group and the corresponding sales data includes: Obtaining multiple financial products purchased by each non-target customer in the customer group and their corresponding original sales data; Filtering the original sales data corresponding to each of the financial products based on a preset second screening rule to obtain valid sales data corresponding to each of the financial products; The cluster analysis of the plurality of financial products based on the purchase tendency data of the target customers and the sales data to obtain a plurality of financial product groups includes: Determining the product characteristics of each of the financial products based on the target customer's purchasing tendency data and the effective sales data; Performing cluster analysis on the plurality of financial products based on the product characteristics to obtain a plurality of financial product groups; Determining the product features of each of the financial products based on the target customer's purchasing tendency data and the sales data includes: Obtaining the target customer's purchasing tendency data, wherein the purchasing tendency data includes long-term holding and trying new products; The product characteristics of each financial product are respectively constructed according to the feature construction rules corresponding to the purchase tendency data and the effective sales data corresponding to each financial product.

2. The personal financial product recommendation method according to claim 1, characterized in that: The cluster analysis is performed on the target customer and a plurality of non-target customers based on their basic information data and financial product purchase data to obtain the customer group to which the target customer belongs, including: Obtaining basic information data and financial product purchase data of the target customer and multiple non-target customers; Filtering the financial product purchase data based on a preset first screening rule to obtain valid financial product purchase data; Build customer characteristics for each customer based on their corresponding basic information data and effective financial product purchase data; Perform cluster analysis on the customers based on the customer characteristics to obtain the customer group to which the target customers belong.

3. The personal financial product recommendation method according to claim 2, characterized in that: After constructing the customer characteristics of each customer based on the basic information data and valid financial product purchase data corresponding to each customer, the method further includes: Screening the customer features based on filtering and recursive feature elimination; Based on the principal component analysis method, feature extraction and feature mapping are performed on each customer feature to obtain the mapped customer features.

4. The personal financial product recommendation method according to claim 1, characterized in that: The recommending financial products to the target customers according to the purchase quantity of each financial product included in each financial product group includes: Obtaining the purchase quantity of each financial product included in each financial product group respectively, and determining the total purchase quantity of each financial product group based on the purchase quantity; Each financial product in the financial product group with the largest total purchase quantity is recommended to the target customer.

5. A personal financial product recommendation device, characterized in that: include: A customer grouping module is used to perform cluster analysis on customers based on the basic information data and financial product purchase data of target customers and multiple non-target customers to obtain the customer group to which the target customers belong; A financial product sales data acquisition module is used to acquire multiple financial products purchased by each non-target customer in the customer group and their corresponding sales data; a financial product grouping module, configured to perform cluster analysis on the plurality of financial products based on the target customer's purchase tendency data and the sales data to obtain a plurality of financial product groups; A financial product recommendation module, configured to recommend financial products to the target customer based on the purchase quantity of each financial product included in each financial product group; The financial product sales data acquisition module includes: a sales data acquisition unit, configured to acquire a plurality of financial products purchased by each non-target customer in the customer group and the corresponding original sales data; a sales data filtering unit, configured to filter the original sales data corresponding to each of the financial products based on a preset second filtering rule to obtain valid sales data corresponding to each of the financial products; The financial product grouping module includes: a financial product feature construction unit, configured to determine the product features of each of the financial products based on the target customer's purchase tendency data and the effective sales data; a financial product grouping unit, configured to perform cluster analysis on the plurality of financial products based on the product characteristics to obtain a plurality of financial product groups; The financial product feature construction unit is specifically used to: Obtaining the target customer's purchasing tendency data, wherein the purchasing tendency data includes long-term holding and trying new products; The product characteristics of each financial product are respectively constructed according to the feature construction rules corresponding to the purchase tendency data and the effective sales data corresponding to each financial product.

6. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the personal financial product recommendation method according to any one of claims 1 to 4 is implemented.

7. A computer-readable storage medium, characterized in that The computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the personal financial product recommendation method according to any one of claims 1 to 4 is implemented.

Images