Financial product recommendation method and device, electronic equipment and storage medium

By determining the neighbor customer set of asset structure similarity and generating a customer-financial product feature matrix, the problem of insufficient accuracy of financial product recommendation caused by sparse data and cold start of users is solved, and more accurate financial product recommendations are achieved.

CN118967262APending Publication Date: 2024-11-15INDUSTRIAL AND COMMERCIAL BANK OF CHINA
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411044279.3
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-07-31
Publication Date
2024-11-15

AI Technical Summary

Technical Problem

In the prior art, due to sparse data and cold start problems of users, the accuracy of the financial product recommendation algorithm is insufficient, which cannot effectively solve the selective difficulties of customers when choosing financial products.

Method used

By determining the neighbor customer set whose asset structure similarity is greater than the preset similarity threshold, obtaining its asset information, generating a customer-financial product feature matrix, and then determining the recommended product result set, using the purchase status of neighbor customers as supplementary data. Improve recommendation accuracy.

Benefits of technology

It improves the accuracy of financial product recommendations, solves the problems of insufficient recommendation accuracy caused by sparse data and cold start of users, and enhances the convenience of customers to choose financial product.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118967262A_ABST
    Figure CN118967262A_ABST
Patent Text Reader

Abstract

The invention discloses a financial product recommendation method and device, electronic equipment and a storage medium, and relates to the field of financial science and technology or other related fields, and the method comprises the steps: determining a neighbor set corresponding to a target customer, the neighbor customers in the neighbor set refer to customers whose asset structure similarity with the target customer is greater than a preset similarity threshold; asset information of the target customer and each neighbor customer in the neighbor set is obtained, a customer asset information set is obtained, and the asset information at least comprises all financial products purchased by the customer in a historical time period and the purchase amount of each financial product; and generating a customer-financial product feature matrix based on the customer asset information set, and determining a recommended product result set associated with the target customer based on the customer-financial product feature matrix. According to the method and the device, the technical problem that accuracy is insufficient when a recommendation algorithm is used for carrying out financial product recommendation due to data sparsity and user cold start in related technologies is solved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of financial technology or other related fields, and in particular to a method for recommending financial products and a device, electronic device, and storage medium thereof. Background Art

[0002] As the concept of financial management in society gradually strengthens, many people will choose to use their idle money to purchase financial products. Currently, there are many types of financial products on the market. The process of selecting financial products is difficult to select, and communication with financial advisors also requires a lot of time and energy. Moreover, most customers do not have large amounts of funds, and usually they will not choose financial products through financial advisors, but will choose financial products independently through different APPs.

[0003] Among the related technologies, the collaborative recommendation algorithm is the most widely used recommendation algorithm in the current recommendation system, but it is prone to problems such as data sparsity and user cold start, which reduces the accuracy of similarity calculation between users and affects the recommendation performance.

[0004] To address the above-mentioned problems, no effective solution has been proposed yet. Summary of the invention

[0005] The embodiments of the present invention provide a method for recommending financial products and a device, electronic device, and storage medium thereof, so as to at least solve the technical problem in the related art that due to data sparsity and user cold start, the accuracy of financial products recommendation using recommendation algorithms is insufficient.

[0006] According to one aspect of an embodiment of the present invention, a method for recommending financial products is provided, comprising: determining a neighbor set corresponding to a target customer, wherein the neighbor set includes M neighbor customers, the neighbor customer refers to a customer whose asset structure similarity with the target customer is greater than a preset similarity threshold, and M is a positive integer; acquiring asset information of the target customer and each of the neighbor customers in the neighbor set to obtain a customer asset information set, wherein the asset information at least includes all financial products purchased by the customer in a historical time period and the purchase amount of each of the financial products; generating a customer-financial product feature matrix based on the customer asset information set; and determining a recommended product result set associated with the target customer based on the customer-financial product feature matrix.

[0007] Optionally, the step of generating a customer-financial product feature matrix based on the customer asset information set includes: traversing all the asset information in the customer asset information set to obtain a union of financial products, wherein the union of financial products includes all the financial products purchased by the target customer and all the neighboring customers within the historical time period; for each customer, determining the customer's product purchase feature vector based on the purchase amount of all the financial products contained in the union of financial products and the asset information; and integrating the product purchase feature vectors corresponding to the target customer and all the neighboring customers to obtain the customer-financial product feature matrix.

[0008] Optionally, the step of determining a recommended product result set associated with the target customer based on the customer-financial product feature matrix includes: determining all unpurchased financial products of the target customer based on the product purchase feature vector of the target customer and the union of financial products to obtain a set of unpurchased financial products; for each of the unpurchased financial products, calculating a first predicted purchase amount and a second predicted purchase amount of the unpurchased financial product by the target customer based on the customer-financial product feature matrix, wherein the first predicted purchase amount refers to a predicted purchase amount calculated based on product purchase characteristics, and the second predicted purchase amount refers to a predicted purchase amount calculated based on product yield characteristics; determining a proposed recommended financial product from the set of unpurchased financial products based on the first predicted purchase amount and the second predicted purchase amount to obtain the recommended product result set.

[0009] Optionally, the step of determining a proposed recommended financial product from the set of unpurchased financial products based on the first predicted purchase amount and the second predicted purchase amount to obtain the recommended product result set includes: if the second predicted purchase amount of the unpurchased financial product is greater than or equal to a preset purchase amount threshold and less than the first predicted purchase amount, storing the unpurchased financial product in the recommended product result set.

[0010] Optionally, for each of the unpurchased wealth management products, the step of calculating the first predicted purchase amount of the target customer for the unpurchased wealth management product based on the customer-wealth management product feature matrix includes: for each of the neighboring customers, extracting the product purchase feature vectors of the neighboring customer and the target customer from the customer-wealth management product feature matrix; calculating the product purchase similarity between the neighboring customer and the target customer based on the product purchase feature vectors of the neighboring customer and the target customer; for each of the unpurchased wealth management products, calculating the average purchase amount of the unpurchased wealth management product by all of the neighboring customers of the target customer; and calculating the first predicted purchase amount of the target customer for the unpurchased wealth management product based on the first prediction formula, the product purchase similarities corresponding to all the neighboring customers and the average purchase amount.

[0011] Optionally, for each of the unpurchased wealth management products, the step of calculating the second predicted purchase amount of the target customer for the unpurchased wealth management product based on the customer-wealth management product feature matrix includes: for each of the wealth management products in the wealth management product set, extracting the product yield feature vector of the wealth management product from the customer-wealth management product feature matrix; for each of the unpurchased wealth management products, calculating the product yield similarity between the unpurchased wealth management product and other wealth management products in the wealth management product set based on the product yield feature vector; and calculating the second predicted purchase amount of the target customer for the unpurchased wealth management product based on the second prediction formula, the product yield similarities corresponding to all the neighboring customers and the average purchase amount.

[0012] Optionally, the step of determining a neighbor set corresponding to a target customer comprises: analyzing the asset structure of the target customer based on an asset database of the financial system to obtain an asset structure indicator, wherein the asset structure indicator is used to describe the asset distribution of the target customer, and the asset structure indicator comprises: asset proportion, risk indicator of an investment portfolio, and rate of return; calculating the asset structure similarity between the target customer and other customers in the financial system based on the asset structure indicator and the asset database; when the asset structure similarity is greater than a preset similarity threshold, determining the other customers corresponding to the asset structure similarity as the neighbor customers of the target customer; and integrating all the neighbor customers of the target customer to obtain the neighbor set.

[0013] According to another aspect of an embodiment of the present invention, a device for recommending financial products is also provided, comprising: a first determination unit, used to determine a neighbor set corresponding to a target customer, wherein the neighbor set includes M neighbor customers, and the neighbor customer refers to a customer whose asset structure similarity with the target customer is greater than a preset similarity threshold, and M is a positive integer; an acquisition unit, used to acquire asset information of the target customer and each of the neighbor customers in the neighbor set to obtain a customer asset information set, wherein the asset information at least includes all financial products purchased by the customer in a historical time period and the purchase amount of each of the financial products; a generation unit, used to generate a customer-financial product feature matrix based on the customer asset information set; and a second determination unit, used to determine a recommended product result set associated with the target customer based on the customer-financial product feature matrix.

[0014] Optionally, the generating unit includes: a traversal module, used to traverse all the asset information in the customer asset information set to obtain a union of financial products, wherein the union of financial products includes all the financial products purchased by the target customer and all the neighboring customers within the historical time period; a first determination module, used to determine, for each customer, a product purchase feature vector of the customer based on the union of financial products and the purchase amount of all the financial products contained in the asset information; and a first integration module, used to integrate the product purchase feature vectors corresponding to the target customer and all the neighboring customers to obtain the customer-financial product feature matrix.

[0015] Optionally, the second determination unit includes: a second determination module, used to determine all unpurchased wealth management products of the target customer based on the product purchase feature vector of the target customer and the union of wealth management products, and obtain a set of unpurchased wealth management products; a first calculation module, used to calculate, for each of the unpurchased wealth management products, a first predicted purchase amount and a second predicted purchase amount of the target customer for the unpurchased wealth management product based on the customer-wealth management product feature matrix, wherein the first predicted purchase amount refers to the predicted purchase amount calculated based on the product purchase characteristics, and the second predicted purchase amount refers to the predicted purchase amount calculated based on the product yield characteristics; a third determination module, used to determine the proposed recommended wealth management product in the set of unpurchased wealth management products based on the first predicted purchase amount and the second predicted purchase amount, and obtain the recommended product result set.

[0016] Optionally, the third determination module includes: a storage submodule, used to store the unpurchased financial product into the recommended product result set when the second predicted purchase amount of the unpurchased financial product is greater than or equal to a preset purchase amount threshold and less than the first predicted purchase amount.

[0017] Optionally, the first calculation module includes: a first extraction submodule, used to extract, for each of the neighbor customers, the product purchase feature vectors of the neighbor customers and the target customers in the customer-financial product feature matrix; a first calculation submodule, used to calculate the product purchase similarity between the neighbor customer and the target customer based on the product purchase feature vectors of the neighbor customers and the target customer; a second calculation submodule, used to calculate, for each of the unpurchased financial products, the average purchase amount of the unpurchased financial products by all of the neighbor customers of the target customer; and a third calculation submodule, used to calculate the first predicted purchase amount of the target customer for the unpurchased financial products based on the first prediction formula, the product purchase similarities corresponding to all the neighbor customers and the average purchase amount.

[0018] Optionally, the first calculation module also includes: a second extraction submodule, used to extract the product yield feature vector of the wealth management product in the customer-wealth management product feature matrix for each wealth management product in the wealth management product set; a fourth calculation submodule, used to calculate, for each of the unpurchased wealth management products, the product yield similarity between the unpurchased wealth management product and other wealth management products in the wealth management product set based on the product yield feature vector; and a fifth calculation submodule, used to calculate the second predicted purchase amount of the target customer for the unpurchased wealth management product based on the second prediction formula, the product yield similarities corresponding to all the neighboring customers and the average purchase amount.

[0019] Optionally, the first determination unit includes: an analysis module, used to analyze the asset structure of the target customer based on the asset database of the financial system to obtain an asset structure indicator, wherein the asset structure indicator is used to describe the asset distribution of the target customer, and the asset structure indicator includes: asset proportion, risk indicator of the investment portfolio and rate of return; a second calculation module, used to calculate the asset structure similarity between the target customer and other customers in the financial system based on the asset structure indicator and the asset database; a fourth determination module, used to determine the other customers corresponding to the asset structure similarity as the neighbor customers of the target customer when the asset structure similarity is greater than the preset similarity threshold; a second integration module, used to integrate all the neighbor customers of the target customer to obtain the neighbor set.

[0020] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is further provided, wherein the computer-readable storage medium includes a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute any one of the above-mentioned methods for recommending financial products.

[0021] According to another aspect of an embodiment of the present invention, there is also provided an electronic device, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement any one of the above-described methods for recommending financial products.

[0022] In the present invention, a method for recommending financial products is proposed. First, a neighbor set corresponding to a target customer is determined, wherein the neighbor set includes M neighbor customers, and a neighbor customer refers to a customer whose asset structure similarity with the target customer is greater than a preset similarity threshold, and M is a positive integer. Then, the asset information of the target customer and each neighbor customer in the neighbor set is obtained to obtain a customer asset information set, wherein the asset information at least includes all financial products purchased by the customer in a historical time period and the purchase amount of each financial product. Then, a customer-financial product feature matrix is ​​generated based on the customer asset information set, and finally a recommended product result set associated with the target customer is determined based on the customer-financial product feature matrix.

[0023] In the present invention, a collaborative filtering algorithm and a recommendation algorithm are combined, and neighbor customers with similar asset structures to the target customer are determined to form a neighbor set, and the asset information of all neighbor customers in the neighbor set in the historical time period, that is, the purchased financial products and the purchase amount, are collected to obtain a customer asset information set, and a customer-financial product feature matrix is ​​generated based on all asset information in the customer asset information set, and then a recommended product result set for the associated target customer is determined. The present invention can determine the neighbor customers of the target customer by the similarity of the customer asset structure, and use the purchase of financial products by the neighbor customers as supplementary data when the original data of the target customer is insufficient. By collecting and analyzing the asset information of the target customer and the neighbor customers of the target customer, the customer-financial product feature matrix is ​​determined, and the recommended financial products for the associated target customer can be determined, and an accurate recommended product result set is obtained, thereby solving the technical problem in the related art that the accuracy is insufficient when using the recommendation algorithm to recommend financial products due to data sparsity and user cold start. BRIEF DESCRIPTION OF THE DRAWINGS

[0024] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0025] Figure 1 is a flow chart of an optional method for recommending financial products according to an embodiment of the present invention;

[0026] Figure 2is a flow chart of an optional financial product recommendation algorithm combining financial product returns and collaborative filtering algorithm according to an embodiment of the present invention;

[0027] Figure 3 is a flow chart of a method for preprocessing customer information and product information according to an embodiment of the present invention;

[0028] Figure 4 is a flowchart of an optional method for calculating a recommendation result set based on asset information according to an embodiment of the present invention;

[0029] Figure 5 is a schematic diagram of an optional device for recommending financial products according to an embodiment of the present invention;

[0030] Figure 6 The present invention is a hardware structure block diagram of an electronic device (or mobile device) for recommending a financial product according to an embodiment of the present invention. DETAILED DESCRIPTION

[0031] In order to enable those skilled in the art to better understand the scheme of the present invention, the technical scheme in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0032] It should be noted that the terms "first", "second", etc. in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments of the present invention described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0033] It should be noted that the method for recommending financial products and the device thereof in the present invention can be used in the field of financial technology when recommending financial products, and can also be used in any field other than the field of financial technology when recommending financial products. The application field of the method for recommending financial products and the device thereof in the present invention is not limited.

[0034] It should be noted that the relevant information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in the present invention are all information and data authorized by the user or fully authorized by all parties, and the collection, storage, processing, transmission, provision, disclosure, use and processing of relevant data need to comply with the laws, regulations and standards of the relevant regions, take necessary confidentiality measures, do not violate public order and good customs, and provide corresponding operation entrances for users to choose to authorize or refuse. For example, an interface is set between this system and relevant users or organizations. Before obtaining relevant information, it is necessary to send an acquisition request to the aforementioned user or organization through the interface, and obtain relevant information after receiving the consent information fed back by the aforementioned user or organization.

[0035] The information collection (for example, user voice, video, and text collection) and analysis operations involved in the present invention have provided users with corresponding operation entrances when they are executed, allowing users to choose to agree or reject the automated decision results; if the user chooses to reject, the expert decision-making process will be entered.

[0036] The following embodiments of the present invention can be applied to various systems / applications / devices that need to recommend financial products, and can use the purchase of financial products by neighboring customers as supplementary data when the original data of the target customer is insufficient in the collaborative recommendation algorithm. The present invention determines the neighboring customers of the target customer by the similarity of the customer's asset structure, collects and analyzes the asset information of the target customer and the neighboring customers of the target customer, determines the customer-financial product feature matrix, and can determine the recommended financial products for the associated target customer, thereby obtaining an accurate recommended product result set.

[0037] The present invention is described in detail below in conjunction with various embodiments.

[0038] Embodiment 1

[0039] According to an embodiment of the present invention, an embodiment of a method for recommending financial products is provided. It should be noted that the steps shown in the flowchart of the accompanying drawings can be executed in a computer system such as a set of computer executable instructions, and although a logical order is shown in the flowchart, in some cases, the steps shown or described can be executed in an order different from that shown here.

[0040] Figure 1 is a flow chart of an optional method for recommending financial products according to an embodiment of the present invention. Figure 1 As shown, the method comprises the following steps:

[0041] Step S101, determining a neighbor set corresponding to a target customer, wherein the neighbor set includes M neighbor customers, and a neighbor customer refers to a customer whose asset structure similarity with the target customer is greater than a preset similarity threshold, and M is a positive integer.

[0042] Step S102, obtaining the asset information of the target customer and each neighbor customer in the neighbor set, and obtaining a customer asset information set, wherein the asset information at least includes all the financial products purchased by the customer in the historical time period and the purchase amount of each financial product.

[0043] Step S103, generating a customer-financial product feature matrix based on the customer asset information set.

[0044] Step S104, determining a recommended product result set associated with the target customer based on the customer-financial product feature matrix.

[0045] Through the above steps, the neighbor set corresponding to the target customer can be determined first, wherein the neighbor set includes M neighbor customers, and the neighbor customer refers to a customer whose asset structure similarity with the target customer is greater than a preset similarity threshold, and M is a positive integer. Then, the asset information of the target customer and each neighbor customer in the neighbor set is obtained to obtain a customer asset information set, wherein the asset information at least includes all the financial products purchased by the customer in the historical time period and the purchase amount of each financial product. Then, a customer-financial product feature matrix is ​​generated based on the customer asset information set, and finally, a recommended product result set associated with the target customer is determined based on the customer-financial product feature matrix.

[0046] In an embodiment of the present invention, a collaborative filtering algorithm and a recommendation algorithm are combined to form a neighbor set by determining neighbor customers with similar asset structures to the target customer, and to collect asset information of all neighbor customers in the neighbor set within a historical time period, that is, the financial products purchased and the purchase amount, to obtain a customer asset information set, and to generate a customer-financial product feature matrix based on all asset information in the customer asset information set, and then to determine a recommended product result set for the associated target customer. The present invention can determine the neighbor customers of the target customer by the similarity of the customer asset structure, and use the purchase of financial products by the neighbor customers as supplementary data when the original data of the target customer is insufficient. By collecting and analyzing the asset information of the target customer and the neighbor customers of the target customer, and determining the customer-financial product feature matrix, the recommended financial products for the associated target customer can be determined, and an accurate recommended product result set can be obtained, thereby solving the technical problem in the related art of insufficient accuracy when using a recommendation algorithm to recommend financial products due to data sparsity and user cold start.

[0047] The embodiment of the present invention is described in detail below in combination with the above steps.

[0048] The implementing body of the embodiment of the present invention can be a recommendation system for financial products. For target customers with less financial data, the financial data of customers with similar asset structures are used as supplementary data, which is used as basic data for use in the financial product recommendation process, thereby solving the problem of insufficient recommendation accuracy caused by data sparsity and user cold start.

[0049] The financial products in the embodiment of the present invention may include: time deposits, money market funds, bond funds, stock funds, ETFs (Exchange Traded Funds), structured deposits, insurance financial products, gold financial products, etc.

[0050] Step S101, determining a neighbor set corresponding to a target customer.

[0051] It should be noted that the neighbor set includes M neighbor customers, and the neighbor customer refers to a customer whose asset structure similarity with the target customer is greater than a preset similarity threshold, and M is a positive integer.

[0052] Optionally, step S101 includes: analyzing the asset structure of the target customer based on the asset database of the financial system to obtain an asset structure indicator, wherein the asset structure indicator is used to describe the asset distribution of the target customer, and the asset structure indicator includes: asset proportion, risk indicator of the investment portfolio, and rate of return; based on the asset structure indicator and the asset database, calculating the asset structure similarity between the target customer and other customers in the financial system; when the asset structure similarity is greater than a preset similarity threshold, determining other customers corresponding to the asset structure similarity as neighbor customers of the target customer; integrating all neighbor customers of the target customer to obtain a neighbor set.

[0053] It should be noted that the asset ratio refers to the proportion of each type of wealth management product in the total assets obtained by dividing the purchase amount of each type of wealth management product by the total asset value; the risk indicator of the investment portfolio refers to the use of the historical volatility and correlation of various types of wealth management products to evaluate the risk level of the wealth management product purchase portfolio. Commonly used risk indicators include standard deviation, Sharpe ratio, beta coefficient, etc.; the rate of return of the investment portfolio refers to the expected rate of return of the customer's overall wealth management product purchase portfolio calculated based on the historical rate of return or expected rate of return of various types of wealth management products, which is used to evaluate the profitability of the wealth management product purchase portfolio.

[0054] It should be noted that in addition to asset ratios, portfolio risk indicators and rates of return, asset structure indicators can also include: asset-liability ratio, asset type distribution, investment term structure, income sources and risk tolerance, etc.

[0055] The embodiments of the present invention may use statistical analysis methods or machine learning algorithms to calculate the asset structure similarity between different customers by comparing the asset proportions, investment portfolio risk indicators and rates of return of different customers. The asset structure similarity may also be obtained by weighting these asset structure indicators using a preset weight ratio. The asset structure similarity between two or more selected customers may also be calculated using a calculation interface provided in the asset database of the financial system. The embodiments of the present invention do not limit the specific calculation method.

[0056] Step S102, obtaining the asset information of the target customer and each neighbor customer in the neighbor set, and obtaining a customer asset information set.

[0057] It should be noted that the asset information at least includes all the financial products purchased by the customer during the historical period and the purchase amount of each financial product.

[0058] It should be noted that asset information may also include: the number of products currently held, the current market value of wealth management products, investment income (including investment yield, realized income, unrealized income, etc.), asset allocation ratio, portfolio risk indicators (volatility, Sharpe ratio, maximum drawdown, etc.), cash flow (including cash flow events such as maturity of time deposits and redemption of wealth management products).

[0059] Step S103, generating a customer-financial product feature matrix based on the customer asset information set.

[0060] Optionally, step S103 includes: traversing all asset information in the customer asset information set to obtain a union of wealth management products, wherein the union of wealth management products includes all wealth management products purchased by the target customer and all neighboring customers in a historical time period; for each customer, determining the customer's product purchase feature vector based on the wealth management product union and the purchase amount of all wealth management products contained in the asset information; integrating the product purchase feature vectors corresponding to the target customer and all neighboring customers to obtain a customer-wealth management product feature matrix.

[0061] For example, the wealth management product set includes n wealth management products (defined as I1, I2, ...Ij...In, n is a positive integer). The product purchase feature vector constructed with the purchase amount of User1 (i.e., the target customer) for these n wealth management products as vector elements is [R11, R12, ...R1j...R1n], where R11 refers to the purchase amount of I1 by User1, and similarly, R1n refers to the purchase amount of In by User1.

[0062] The embodiment of the present invention can determine the product purchase feature vector of each neighbor customer according to the above feature vector construction rule, and integrate the product purchase feature vectors of all customers into a customer-financial product feature matrix.

[0063] In the process of integrating the customer-financial product feature matrix, the embodiment of the present invention can use all product purchase feature vectors as the row vectors of the matrix, and the obtained customer-financial product feature matrix is ​​exemplified as follows:

[0064]

[0065] Where j is a positive integer less than n, R m1 It refers to the purchase amount of I1 by Userm, Rmn refers to the purchase amount of In by Userm, i is a positive integer less than m, and the column vector of the customer-financial product feature matrix obtained in this way is the product yield feature vector of each financial product.

[0066] In addition, you can also choose to use all product purchase feature vectors as the column vectors of the matrix, so that the row vectors of the customer-financial product feature matrix obtained are the product return feature vectors of each financial product.

[0067] Step S104, determining a recommended product result set associated with the target customer based on the customer-financial product feature matrix.

[0068] Optionally, step S104 includes: based on the target customer's product purchase feature vector and the union of wealth management products, determining all unpurchased wealth management products of the target customer to obtain a set of unpurchased wealth management products; for each unpurchased wealth management product, calculating a first predicted purchase amount and a second predicted purchase amount of the target customer for the unpurchased wealth management product based on the customer-wealth management product feature matrix, wherein the first predicted purchase amount refers to a predicted purchase amount calculated based on the product purchase characteristics, and the second predicted purchase amount refers to a predicted purchase amount calculated based on the product yield characteristics; based on the first predicted purchase amount and the second predicted purchase amount, determining a wealth management product to be recommended in the set of unpurchased wealth management products to obtain a recommended product result set.

[0069] Continuing with the above example, for User1 (i.e., the target customer), the product purchase features can be represented by the product purchase feature vector, i.e., [R11, R12, …R1j…R1n], where n represents the financial product and the set includes n financial products (defined as I1, I2, …Ij…In, n is a positive integer), R11 refers to the purchase amount of I1 by User1, and similarly, R1n refers to the purchase amount of In by User1.

[0070] The product purchase features are extracted through the customer-financial product feature matrix, which can be represented as [R11, R21, …Ri1…Rm1], where m means there are a total of m target customers and neighbor customers (defined as User1, User2, …Userj…Userm, m is a positive integer), R11 refers to the purchase amount of I1 by User1, and similarly, Rm1 refers to the purchase amount of I1 by Userm.

[0071] Optionally, the step of determining the proposed recommended financial product in the set of unpurchased financial products based on the first predicted purchase amount and the second predicted purchase amount to obtain a recommended product result set includes: when the second predicted purchase amount of the unpurchased financial product is greater than or equal to a preset purchase amount threshold and less than the first predicted purchase amount, storing the unpurchased financial product in the recommended product result set.

[0072] Optionally, for each unpurchased wealth management product, the step of calculating the first predicted purchase amount of the target customer for the unpurchased wealth management product based on the customer-wealth management product feature matrix includes: for each neighbor customer, extracting the product purchase feature vectors of the neighbor customer and the target customer from the customer-wealth management product feature matrix; calculating the product purchase similarity between the neighbor customer and the target customer based on the product purchase feature vectors of the neighbor customer and the target customer; for each unpurchased wealth management product, calculating the average purchase amount of the unpurchased wealth management product by all neighbor customers of the target customer; and calculating the first predicted purchase amount of the target customer for the unpurchased wealth management product based on the first prediction formula, the product purchase similarities corresponding to all neighbor customers and the average purchase amount.

[0073] It should be noted that the product purchase feature vector of target customer u is extracted and the product purchase feature vector of neighbor customer v By calculating the vector and The cosine angle between is used to characterize the product purchase similarity between target customer u and neighbor customer v. For example, Wherein, sim(u,v) is the product purchase similarity. The embodiment of the present invention provides an exemplary first prediction formula: Among them, P u,i It refers to the first predicted purchase amount of target customer u for the financial product i that he has not purchased. It refers to the average purchase amount of product i by u’s neighbors, N u represents the neighbor set of target customer u, r v,i N u The purchase amount of product i by neighbor customer v in .

[0074] In addition, after using cosine similarity to calculate product purchase similarity, the embodiment of the present invention can also use the Pearson correlation coefficient to refine the similarity calculation formula. The Pearson correlation coefficient assumes that there is a continuous purchase amount range and a linear relationship, normalizes the purchase amount scale, and all purchase amount values ​​are not empty. The similarity value is [0, 1].

[0075] The refined similarity calculation formula is: Among them, I uv represents the union of the financial products of target customer u and neighbor customer v; R u,i and R v,i Respectively represent the purchase amount of financial product i by target customer u and neighbor customer v; and Represent the average purchase amount of target customer u and neighbor customer v respectively.

[0076] Optionally, for each unpurchased wealth management product, the step of calculating the second predicted purchase amount of the target customer for the unpurchased wealth management product based on the customer-wealth management product feature matrix includes: for each wealth management product in the wealth management product set, extracting the product yield feature vector of the wealth management product from the customer-wealth management product feature matrix; for each unpurchased wealth management product, calculating the product yield similarity between the unpurchased wealth management product and other wealth management products in the wealth management product set based on the product yield feature vector; and calculating the second predicted purchase amount of the target customer for the unpurchased wealth management product based on the second prediction formula, the product yield similarities corresponding to all neighboring customers and the average purchase amount.

[0077] It should be noted that the product income feature vectors of financial products i and j are extracted and By calculating the vector and The cosine angle between is used to characterize the product return similarity between financial product i and financial product j. For example, Among them, sim(i,j) is the product revenue similarity.

[0078] The embodiment of the present invention provides an exemplary second prediction formula: Among them, R u,i It refers to the second predicted purchase amount of target customer u for the financial product i that he has not purchased. It refers to the average purchase amount of product i by u’s neighbors, N i represents the product set consisting of all other financial products purchased by all neighboring customers who have purchased product i, r u,j is the target customer u to N i The purchase amount of other financial products j in .

[0079] In the embodiment of the present invention, the data processing method of collaborative filtering, vector error correction and recommendation algorithm is used to process the user's purchase data of financial products to obtain a recommendation list for the user. Compared with the common filtering algorithm, the financial product recommendation method provided in this embodiment can avoid the problems of user cold start and data sparseness.

[0080] The present invention is described below in conjunction with another specific implementation manner.

[0081] The embodiment of the present invention provides a financial product recommendation algorithm that combines financial product returns with a collaborative filtering algorithm, and is used to improve the accuracy and success rate of financial institutions recommending financial products to target customers.

[0082] Figure 2 is a flowchart of an optional financial product recommendation algorithm combining financial product returns and collaborative filtering algorithm according to an embodiment of the present invention, such as Figure 2 As shown, the algorithm includes the following steps:

[0083] S201, pre-processing the information related to the customer's assets to screen out the characteristics of the asset information, and pre-processing the recent yield curve information of the wealth management product to screen out the characteristics of the yield curve;

[0084] S202, inputting user asset information to obtain a recommendation result set based on the asset information;

[0085] S203, inputting the yield curve characteristics of the financial product to obtain a recommendation result set based on the yield curve of the financial product;

[0086] S204, using the recommended result set of asset information as a weight, processing the recommended result set of the yield curve of the financial product.

[0087] Figure 3 is a flow chart of a method for preprocessing customer information and product information according to an embodiment of the present invention. Figure 3 As shown, the method comprises the following steps:

[0088] S2011, obtain customer asset information data, the data comes from the data collected by the financial system, including deposit information, personal information, income information and loan information.

[0089] S2012, collect customer asset information, group users with similar asset structures together, and determine the user's neighbor set.

[0090] S2013, collects information on the yield curve of financial products. The data comes from the data collected by the financial system, including the starting point of purchase, yield, number of buyers, purchase amount and term of the financial products.

[0091] S2014, processing the yield curve information of the financial product, and making the user information and the purchase amount of the financial product into an m*n matrix.

[0092] As shown in Table 1, the m×n matrix R mn Represents user purchase amount data, where m rows represent m users, n columns represent n financial products, and the element R in the i-th row and j-th column in the matrix represents the purchase amount of financial product j by user i.

[0093] Table 1

[0094]

[0095] Figure 4 is a flowchart of an optional method for calculating a recommendation result set based on asset information according to an embodiment of the present invention. Figure 4 As shown, the method comprises the following steps:

[0096] S2021, calculate the similarity between user u and user v based on the asset information.

[0097] Consider the purchase amount of user u as a vector in the n-dimensional financial product space The similarity between users can be calculated by the cosine angle between the purchase amount vectors.

[0098]

[0099] Among them, sim(u,v) represents the similarity between user u and user v.

[0100] S2022, using the Pearson correlation coefficient to refine the cosine similarity to obtain a similarity with a smaller error.

[0101] The Pearson correlation coefficient is used to refine formula (1). The Pearson correlation coefficient assumes that there is a continuous purchase amount range and a linear relationship between the variables to normalize the purchase amount scale. Since the valid users already have the purchase amounts of all financial products in the normalized set, all purchase amount values ​​are not empty. The similarity between user u and the final target user v is [0,1].

[0102]

[0103] In the formula, I uv represents the common purchase items of user u and user v; R u,i and R v,i Respectively represent the purchase amount of financial product i by user u and user v; and Represent the average purchase amount of user u and user v respectively.

[0104] S2023, reducing the calculation error of cosine similarity by eliminating the average value of the purchase amount of the user's financial products.

[0105]

[0106] Among them, I uv I represents the set of financial products purchased by user u and user v together; u and I v Represent the sets of financial products purchased by user u and user v respectively.

[0107] S2024, predicting the purchase amount of unpurchased financial products by the target user.

[0108]

[0109] Among them, N u represents the neighbor set of target user u, Represents the average purchase amount of u’s neighbors.

[0110] In the embodiment of the present invention, the data processing method of collaborative filtering, vector error correction and recommendation algorithm is used to process the user's purchase data of financial products to obtain a recommendation list for the user. Compared with ordinary filtering algorithms, this method can avoid the problems of user cold start and data sparseness.

[0111] The present invention is described below in conjunction with another optional embodiment.

[0112] Embodiment 2

[0113] A financial product recommendation device provided in this embodiment includes multiple implementation units, each of which corresponds to each implementation step in the above-mentioned embodiment 1.

[0114] Figure 5 is a schematic diagram of an optional device for recommending financial products according to an embodiment of the present invention, such as Figure 5 As shown, the device may include: a first determining unit 51, an acquiring unit 52, a generating unit 53, and a second determining unit 54.

[0115] The first determining unit 51 is used to determine a neighbor set corresponding to the target customer, wherein the neighbor set includes M neighbor customers, and the neighbor customer refers to a customer whose asset structure similarity with the target customer is greater than a preset similarity threshold, and M is a positive integer.

[0116] The acquisition unit 52 is used to acquire the asset information of the target customer and each neighbor customer in the neighbor set to obtain a customer asset information set, wherein the asset information at least includes all the financial products purchased by the customer in the historical time period and the purchase amount of each financial product.

[0117] The generating unit 53 is used to generate a customer-financial product characteristic matrix based on the customer asset information set.

[0118] The second determining unit 54 is used to determine a recommended product result set associated with the target customer based on the customer-financial product feature matrix.

[0119] The above-mentioned financial product recommendation device can first determine the neighbor set corresponding to the target customer through the first determination unit 51, wherein the neighbor set includes M neighbor customers, and the neighbor customer refers to a customer whose asset structure similarity with the target customer is greater than a preset similarity threshold, and M is a positive integer. Then, the asset information of the target customer and each neighbor customer in the neighbor set is obtained through the acquisition unit 52 to obtain a customer asset information set, wherein the asset information at least includes all financial products purchased by the customer in a historical time period and the purchase amount of each financial product. Then, the customer-financial product feature matrix is ​​generated based on the customer asset information set through the generation unit 53, and finally, the recommended product result set associated with the target customer is determined based on the customer-financial product feature matrix through the second determination unit 54.

[0120] In an embodiment of the present invention, a collaborative filtering algorithm and a recommendation algorithm are combined to form a neighbor set by determining neighbor customers with similar asset structures to the target customer, and to collect asset information of all neighbor customers in the neighbor set within a historical time period, that is, the financial products purchased and the purchase amount, to obtain a customer asset information set, and to generate a customer-financial product feature matrix based on all asset information in the customer asset information set, and then to determine a recommended product result set for the associated target customer. The present invention can determine the neighbor customers of the target customer by the similarity of the customer asset structure, and use the purchase of financial products by the neighbor customers as supplementary data when the original data of the target customer is insufficient. By collecting and analyzing the asset information of the target customer and the neighbor customers of the target customer, and determining the customer-financial product feature matrix, the recommended financial products for the associated target customer can be determined, and an accurate recommended product result set can be obtained, thereby solving the technical problem of insufficient accuracy when using a recommendation algorithm to recommend financial products due to data sparsity and user cold start in the related technology.

[0121] Optionally, the generation unit includes: a traversal module, used to traverse all asset information in the customer asset information set to obtain a union of financial products, wherein the union of financial products includes all financial products purchased by the target customer and all neighboring customers within a historical time period; a first determination module, used to determine the product purchase feature vector of each customer based on the purchase amount of all financial products by the customer contained in the union of financial products and the asset information; and a first integration module, used to integrate the product purchase feature vectors corresponding to the target customer and all neighboring customers to obtain a customer-financial product feature matrix.

[0122] Optionally, the second determination unit includes: a second determination module, used to determine all unpurchased wealth management products of the target customer based on the target customer's product purchase feature vector and the union of wealth management products, and obtain a set of unpurchased wealth management products; a first calculation module, used to calculate, for each unpurchased wealth management product, a first predicted purchase amount and a second predicted purchase amount of the target customer for the unpurchased wealth management product based on the customer-wealth management product feature matrix, wherein the first predicted purchase amount refers to the predicted purchase amount calculated based on the product purchase characteristics, and the second predicted purchase amount refers to the predicted purchase amount calculated based on the product yield characteristics; a third determination module, used to determine the proposed recommended wealth management product in the set of unpurchased wealth management products based on the first predicted purchase amount and the second predicted purchase amount, and obtain a recommended product result set.

[0123] Optionally, the third determination module includes: a storage submodule, used to store the unpurchased financial product into the recommended product result set when the second predicted purchase amount of the unpurchased financial product is greater than or equal to a preset purchase amount threshold and less than the first predicted purchase amount.

[0124] Optionally, the first calculation module includes: a first extraction submodule, used to extract, for each neighbor customer, the product purchase feature vectors of the neighbor customer and the target customer in the customer-financial product feature matrix; a first calculation submodule, used to calculate the product purchase similarity between the neighbor customer and the target customer based on the product purchase feature vectors of the neighbor customer and the target customer; a second calculation submodule, used to calculate, for each financial product that has not been purchased, the average purchase amount of the target customer's neighbor customers for the unpurchased financial products; and a third calculation submodule, used to calculate the first predicted purchase amount of the target customer for the unpurchased financial products based on the first prediction formula, the product purchase similarities corresponding to all neighbor customers and the average purchase amount.

[0125] Optionally, the first calculation module also includes: a second extraction sub-module, used to extract the product return feature vector of the wealth management product in the customer-wealth management product feature matrix for each wealth management product in the wealth management product set; a fourth calculation sub-module, used to calculate, for each unpurchased wealth management product, the product return similarity between the unpurchased wealth management product and other wealth management products in the wealth management product set based on the product return feature vector; and a fifth calculation sub-module, used to calculate the second predicted purchase amount of the target customer for the unpurchased wealth management product based on the second prediction formula, the product return similarities corresponding to all neighboring customers and the average purchase amount.

[0126] Optionally, the first determination unit includes: an analysis module, which is used to analyze the asset structure of the target customer based on the asset database of the financial system to obtain an asset structure indicator, wherein the asset structure indicator is used to describe the asset distribution of the target customer, and the asset structure indicator includes: asset proportion, risk indicator of the investment portfolio, and rate of return; a second calculation module, which is used to calculate the asset structure similarity between the target customer and other customers in the financial system based on the asset structure indicator and the asset database; a fourth determination module, which is used to determine other customers corresponding to the asset structure similarity as neighbor customers of the target customer when the asset structure similarity is greater than a preset similarity threshold; a second integration module, which is used to integrate all neighbor customers of the target customer to obtain a neighbor set.

[0127] The above-mentioned financial product recommendation device may also include a processor and a memory. The above-mentioned first determination unit 51, acquisition unit 52, generation unit 53, second determination unit 54, etc. are all stored in the memory as program units, and the processor executes the above-mentioned program units stored in the memory to realize corresponding functions.

[0128] The processor includes a kernel, which retrieves the corresponding program unit from the memory. One or more kernels can be set, and the kernel parameters are adjusted to generate a customer-financial product feature matrix based on the customer asset information set, and determine the recommended product result set associated with the target customer based on the customer-financial product feature matrix.

[0129] The above-mentioned memory may include non-permanent memory in a computer-readable medium, random access memory (RAM) and / or non-volatile memory, such as read-only memory (ROM) or flash RAM, and the memory includes at least one storage chip.

[0130] The present application also provides a computer program product, which, when executed on a data processing device, is suitable for executing a program that initializes the following method steps: determining a neighbor set corresponding to a target customer, wherein the neighbor set includes M neighbor customers, and a neighbor customer refers to a customer whose asset structure similarity with the target customer is greater than a preset similarity threshold, and M is a positive integer; obtaining asset information of the target customer and each neighbor customer in the neighbor set to obtain a customer asset information set, wherein the asset information at least includes all financial products purchased by the customer in a historical time period and the purchase amount of each financial product; generating a customer-financial product feature matrix based on the customer asset information set; and determining a recommended product result set associated with the target customer based on the customer-financial product feature matrix.

[0131] According to another aspect of an embodiment of the present invention, a computer-readable storage medium is further provided, the computer-readable storage medium including a stored computer program, wherein when the computer program is running, the device where the computer-readable storage medium is located is controlled to execute the method for recommending financial products of any one of the above-mentioned embodiments one.

[0132] According to another aspect of an embodiment of the present invention, there is also provided an electronic device, comprising one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by one or more processors, the one or more processors implement the method for recommending financial products of any one of the above-mentioned embodiments one.

[0133] Figure 6 1 is a hardware structure block diagram of an electronic device (or mobile device) for a method for recommending financial products according to an embodiment of the present invention. Figure 6 As shown, the electronic device may include one or more ( Figure 6 602a, 602b, ..., 602n are used to illustrate) processor 602 (processor 602 may include but is not limited to a processing device such as a microprocessor MCU or a programmable logic device FPGA), and memory 604 for storing data. In addition, it may also include: a display, an input / output interface (I / O interface), a universal serial bus (USB) port (which may be included as one of the ports of the I / O interface), a network interface, a keyboard, a power supply and / or a camera. It can be understood by those skilled in the art that Figure 6 The structure shown is only for illustration and does not limit the structure of the above electronic device. Figure 6 More or fewer components as shown, or with Figure 6 Different configurations are shown.

[0134] The serial numbers of the above embodiments of the present invention are only for description and do not represent the advantages or disadvantages of the embodiments.

[0135] In the above embodiments of the present invention, the description of each embodiment has its own emphasis. For parts that are not described in detail in a certain embodiment, reference can be made to the relevant descriptions of other embodiments.

[0136] In the several embodiments provided in this application, it should be understood that the disclosed technical content can be implemented in other ways. Among them, the device embodiments described above are only schematic. For example, the division of units can be a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of units or modules, which can be electrical or other forms.

[0137] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple units. Some or all of the units may be selected according to actual needs to achieve the purpose of the present embodiment.

[0138] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0139] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), mobile hard disk, magnetic disk or optical disk and other media that can store program codes.

[0140] The above is only a preferred embodiment of the present invention. It should be pointed out that for ordinary technicians in this technical field, several improvements and modifications can be made without departing from the principle of the present invention. These improvements and modifications should also be regarded as the scope of protection of the present invention.

Claims

1. A method for recommending financial products. It is characterized in that include: Determine a neighbor set corresponding to the target customer, wherein the neighbor set includes M neighbor customers, and the neighbor customer refers to a customer whose asset structure similarity with the target customer is greater than a preset similarity threshold, and M is a positive integer; Acquire the asset information of the target customer and each of the neighbor customers in the neighbor set to obtain a customer asset information set, wherein the asset information at least includes all the financial products purchased by the customer in a historical time period and the purchase amount of each of the financial products; Generate a customer-financial product feature matrix based on the customer asset information set; A recommended product result set associated with the target customer is determined based on the customer-financial product feature matrix.

2. The recommendation method according to claim 1, It is characterized in that The step of generating a customer-financial product feature matrix based on the customer asset information set includes: Traversing all the asset information in the customer asset information set to obtain a union of financial products, wherein the union of financial products includes all the financial products purchased by the target customer and all the neighboring customers within the historical time period; For each customer, determining a product purchase feature vector of the customer based on the union of the financial products and the purchase amount of all the financial products by the customer contained in the asset information; The product purchase feature vectors corresponding to the target customer and all the neighboring customers are integrated to obtain the customer-financial product feature matrix.

3. The recommendation method according to claim 2, It is characterized in that The step of determining a recommended product result set associated with the target customer based on the customer-financial product feature matrix comprises: Based on the product purchase feature vector of the target customer and the union of financial products, determine all unpurchased financial products of the target customer to obtain a set of unpurchased financial products; For each of the unpurchased financial products, a first predicted purchase amount and a second predicted purchase amount of the target customer for the unpurchased financial product are calculated based on the customer-financial product feature matrix, wherein the first predicted purchase amount refers to the predicted purchase amount calculated based on the product purchase features, and the second predicted purchase amount refers to the predicted purchase amount calculated based on the product yield features; Based on the first predicted purchase amount and the second predicted purchase amount, a proposed recommended financial product is determined in the set of unpurchased financial products to obtain the recommended product result set.

4. The recommendation method according to claim 3, It is characterized in that The step of determining a proposed recommended financial product in the set of unpurchased financial products based on the first predicted purchase amount and the second predicted purchase amount to obtain the recommended product result set comprises: When the second predicted purchase amount of the unpurchased financial product is greater than or equal to a preset purchase amount threshold and less than the first predicted purchase amount, the unpurchased financial product is stored in the recommended product result set.

5. The recommendation method according to claim 3, It is characterized in that For each of the unpurchased financial products, the step of calculating the first predicted purchase amount of the target customer for the unpurchased financial product based on the customer-financial product feature matrix comprises: For each of the neighboring customers, extract the product purchase feature vectors of the neighboring customer and the target customer in the customer-financial product feature matrix; Based on the product purchase feature vectors of the neighboring customer and the target customer, calculating the product purchase similarity between the neighboring customer and the target customer; For each of the unpurchased financial products, calculate the average purchase amount of the unpurchased financial products by all of the neighboring customers of the target customer; Based on the first prediction formula, the product purchase similarities corresponding to all the neighboring customers and the average purchase amount, a first predicted purchase amount of the target customer for the unpurchased financial product is calculated.

6. The recommendation method according to claim 3, It is characterized in that For each of the unpurchased financial products, the step of calculating the second predicted purchase amount of the target customer for the unpurchased financial product based on the customer-financial product feature matrix comprises: For each of the wealth management products in the wealth management set, extracting a product yield feature vector of the wealth management product in the customer-wealth management product feature matrix; For each of the unpurchased financial products, calculating the product yield similarity between the unpurchased financial product and other financial products in the wealth management product set based on the product yield feature vector; Based on the second prediction formula, the product return similarity and the average purchase amount corresponding to all the neighboring customers, the second predicted purchase amount of the target customer for the unpurchased financial product is calculated.

7. The recommendation method according to claim 1, It is characterized in that The steps of determining the neighbor set corresponding to the target customer include: Analyzing the asset structure of the target customer based on the asset database of the financial system to obtain an asset structure indicator, wherein the asset structure indicator is used to describe the asset distribution of the target customer, and the asset structure indicator includes: asset ratio, risk indicator of the investment portfolio and rate of return; Calculating the asset structure similarity between the target customer and other customers in the financial system based on the asset structure indicator and the asset database; When the asset structure similarity is greater than the preset similarity threshold, determining the other customers corresponding to the asset structure similarity as the neighboring customers of the target customer; All the neighbor customers of the target customer are integrated to obtain the neighbor set.

8. A device for recommending financial products. It is characterized in that include: A first determining unit is used to determine a neighbor set corresponding to a target customer, wherein the neighbor set includes M neighbor customers, and the neighbor customer refers to a customer whose asset structure similarity with the target customer is greater than a preset similarity threshold, and M is a positive integer; An acquisition unit, configured to acquire asset information of the target customer and each of the neighbor customers in the neighbor set, to obtain a customer asset information set, wherein the asset information at least includes all financial products purchased by the customer in a historical time period and the purchase amount of each of the financial products; A generating unit, configured to generate a customer-financial product feature matrix based on the customer asset information set; The second determining unit is used to determine a recommended product result set associated with the target customer based on the customer-financial product feature matrix.

9. A computer-readable storage medium, It is characterized in that The computer-readable storage medium includes a stored computer program, wherein when the computer program is executed, the device where the computer-readable storage medium is located is controlled to execute the method for recommending financial products according to any one of claims 1 to 7.

10. An electronic device, It is characterized in that It includes one or more processors and a memory, wherein the memory is used to store one or more programs, wherein when the one or more programs are executed by the one or more processors, the one or more processors implement the financial product recommendation method described in any one of claims 1 to 7.