An artificial intelligence-based product pushing method

Abstract

The application discloses a product pushing method based on artificial intelligence, and the application obtains asset information of each bank customer in a target bank, evaluates risk preference estimation coefficients of each bank customer in the target bank corresponding to each risk type, compares and obtains the risk type corresponding to each bank customer in the target bank, thereby realizing data-based, systematic and intelligent analysis on the risk type corresponding to the bank customer, improving the certainty and accuracy of the analysis result of the risk type corresponding to the bank customer, meanwhile, extracting relevant data of each bank customer in the target bank corresponding to each financial product, evaluating recommendation coefficients of each bank customer in the target bank corresponding to each financial product, and sequentially sorting according to the order from high to low of the recommendation coefficients, and pushing to the corresponding bank customer, so that the bank customer can quickly find the financial product meeting the psychological expectation, and the purchase experience of the bank customer is improved.

Description

Technical Field

[0001] This invention relates to the field of intelligent product recommendation technology, and specifically to a product recommendation method based on artificial intelligence. Background Technology

[0002] With economic development and the growth of residents' disposable wealth, people are becoming increasingly enthusiastic about wealth management products, and their demand for them is also increasing. More and more people prefer to purchase wealth management products on banks' online platforms. However, the sheer number of wealth management products available on these platforms makes it difficult for users to choose. Therefore, effectively recommending suitable wealth management products to users has become a new goal for banks' online platforms.

[0003] Currently, online banking platforms use risk assessment questionnaires to determine users' risk preferences before they purchase wealth management products. However, users may be influenced by personal psychological factors when filling out the questionnaire, making the results unable to accurately reflect their risk tolerance. This results in significant uncertainty and inaccuracy, leading to recommended wealth management products that do not match the user's actual risk tolerance. Consequently, the accuracy of wealth management product recommendations on online banking platforms is reduced, failing to meet the wealth management and investment needs of bank users.

[0004] Currently, banks' online platforms recommend wealth management products by pushing all products that match users' risk preferences to the platform's recommended section. However, due to the large number of products, users need to spend a lot of time comparing and reviewing them one by one, making it difficult for them to find products that meet their expectations. This leads to a decreased user experience and further diminishes users' enthusiasm for purchasing wealth management products. Consequently, the sales volume of wealth management products on banks' online platforms cannot increase rapidly, which has a certain impact on the rapid development of banks' online platforms. Summary of the Invention

[0005] In view of this, in order to solve the problems mentioned in the background technology, an artificial intelligence-based product recommendation method is proposed.

[0006] To achieve the above objectives, the present invention provides a product recommendation method based on artificial intelligence, comprising the following steps:

[0007] S1. Obtaining Bank Customer Asset Information: Obtaining asset information of customers of each bank within the target bank, including account information and product holding information;

[0008] S2. Bank Customer Asset Information Analysis: Analyze the asset information of each bank customer within the target bank to obtain the compliance ratio index of each bank customer's corresponding account information and held product information with each risk type.

[0009] S3. Risk Preference Prediction Coefficient Assessment: Based on the compliance ratio index of each risk type for each bank customer's account information and held product information within the target bank, assess the risk preference prediction coefficient for each risk type for each bank customer within the target bank.

[0010] S4. Selection of eligible wealth management products: Compare the risk preference prediction coefficients of customers of each bank within the target bank for each risk type to obtain the risk type of customers of each bank within the target bank, and select eligible wealth management products for customers of each bank within the target bank.

[0011] S5. Data Extraction for Eligible Wealth Management Products: Extract relevant data for each eligible wealth management product corresponding to each bank customer within the target bank. The relevant data includes yield curves, unit net value curves, and management personnel information.

[0012] S6. Evaluation of Recommendation Coefficient for Eligible Wealth Management Products: The evaluation assesses the recommendation coefficient of each eligible wealth management product for each bank customer within the target bank, and sorts them in descending order of recommendation coefficient, while simultaneously pushing the results to the corresponding bank customers.

[0013] Preferably, the account information includes age, gender, average monthly income, and average monthly expenditure, and the product holding information includes the risk type, holding period, holding amount, and rate of return for each product held.

[0014] Preferably, the method for parsing the compliance ratio index of each risk type for the account information corresponding to each bank customer within the target bank in step S2 is as follows:

[0015] Based on the account information of customers in each bank within the target bank, extract the age, gender, average monthly income, and average monthly expenditure of customers in each bank within the target bank. Subtract the average monthly expenditure of the corresponding bank customer from the average monthly income of customers in each bank within the target bank to obtain the average monthly savings amount of customers in each bank within the target bank.

[0016] Extract the risk tolerance weights for each risk type corresponding to customers of different age groups, genders, and average monthly savings amounts stored in the bank's data repository. Based on the age, gender, and average monthly savings amount of customers within the target bank, filter the risk tolerance weights for each risk type corresponding to the age, gender, and average monthly savings amount of customers within the target bank, and label them as p. i c a1、p i c a2、p i c a3, i = 1, 2, ..., n, where i represents the ID of the i-th bank customer, and c = 1, 2, ..., k, where c represents the ID of the c-th risk type;

[0017] Analysis of the compliance ratio index of customer account information for each risk type within the target bank. δ1, δ2, and δ3 represent the risk type compliance factors corresponding to the preset customer age, customer gender, and average monthly savings amount, respectively.

[0018] Preferably, in step S2, the method for analyzing the conformity index of the product information held by each bank customer within the target bank for each risk type is as follows:

[0019] Based on the product holding information of customers within the target bank, extract the risk type, holding period, holding amount, and rate of return for each product held by each customer within the target bank. Count the number of products held by each customer within the target bank for each risk type and label them as x. i c ;

[0020] Extract the holding time, holding amount, and rate of return of each product held by each bank customer within the target bank for each risk type, and obtain the preference of each bank customer for each risk type of held products within the target bank, which is denoted as φ. i c ;

[0021] Analysis of the compliance ratio index of product holdings of customers of each bank within the target bank for each risk type Where μ represents the preset correction factor for customer-held product information, and e represents the natural constant.

[0022] Preferably, the risk preference prediction coefficients for each risk type of each bank customer within the target bank in step S3 are assessed using the following method:

[0023] The compliance ratio index ξ of customer account information of each bank within the target bank for each risk type. i c The compliance ratio index ψ of the product holding information of each bank's customers with each risk type i c Substitute into the risk preference prediction coefficient assessment formula θ i c =λ1*ξ i c +λ2*ψ i c The risk preference prediction coefficients θ for each risk type of customers within the target bank are obtained. i c λ1 and λ2 represent the weighting factors corresponding to the preset customer account information and the weighting factors corresponding to the customer's product information, respectively, and λ1+λ2=1.

[0024] Preferably, the specific steps corresponding to step S4 are as follows:

[0025] Compare the risk preference prediction coefficients of customers of each bank in the target bank for each risk type, and select the risk type with the highest risk preference prediction coefficient for customers of each bank in the target bank, and record it as the risk type for customers of each bank in the target bank.

[0026] Extract the risk types corresponding to each wealth management product from the target bank's platform. Based on the risk types corresponding to each bank's customers within the target bank, filter out the wealth management products whose risk types match those of each bank's customers within the target bank, and record them as the eligible wealth management products corresponding to each bank's customers within the target bank.

[0027] Preferably, step S5 involves extracting relevant data for each eligible wealth management product corresponding to each bank customer within the target bank, specifically including:

[0028] Extract relevant data of each bank's customers corresponding to each eligible wealth management product from the target bank's platform, and obtain the yield curve, unit net value curve and management personnel information of each bank's customers corresponding to each eligible wealth management product within the target bank.

[0029] Based on the yield curves and net asset value curves of each eligible wealth management product for each bank's customers within the target bank, the yields and net asset values ​​of each eligible wealth management product for each bank's customers within the target bank at each data collection point within a preset time period are obtained. The yield volatility index, yield growth index, and net asset value growth index of each eligible wealth management product for each bank's customers within the target bank during the preset time period are analyzed and denoted as α. ir d1, α ir d2, α ir d3, r = 1, 2, ..., u, where r represents the r-th product number that meets the criteria for a financial product;

[0030] Based on the information of the management personnel of each eligible wealth management product for each customer within the target bank, a weighted index for the management personnel information of each eligible wealth management product for each customer within the target bank is obtained and labeled as β. ir .

[0031] Preferably, the formula for analyzing the yield volatility index of each bank customer within the target bank for each eligible wealth management product within a preset time period is as follows: Where α ir d1 represents the yield volatility index of the r-th eligible wealth management product for the i-th bank customer within the target bank during the preset time period, h represents the number of data collection points within the preset time period, and y ir d fLet f represent the yield of the r-th eligible wealth management product for the i-th bank customer within the target bank at the f-th data collection point within the preset time period, where f = 1, 2, ..., h. This represents the allowable standard deviation corresponding to the preset yield of the wealth management product.

[0032] Preferably, the weighted index parsing method for the management personnel information of each bank customer within the target bank corresponding to each eligible wealth management product is as follows:

[0033] Extract the management personnel information for each eligible wealth management product corresponding to each bank customer within the target bank. This information includes the manager's years of wealth management experience, average annualized return, and cumulative managed amount. Label the years of wealth management experience, average annualized return, and cumulative managed amount in the management personnel information for each eligible wealth management product corresponding to each bank customer within the target bank as w. ir g1, w ir g2、w ir g3,

[0034] Extract the average years of wealth management experience, average annualized return, and average cumulative managed amount of wealth management product managers from the target bank platform, and label them as follows:

[0035] Based on the management personnel information, it conforms to the weighted index analysis formula. Information on the management personnel of each eligible wealth management product for each bank's customers within the target bank is obtained, conforming to the weighted index β. ir ε1, ε2, and ε3 represent the default influencing factors corresponding to the management personnel's years of financial management experience, average annualized rate of practice, and cumulative amount of management, respectively.

[0036] Preferably, in step S6, the recommendation coefficients for each eligible wealth management product for each bank customer within the target bank are evaluated, and the specific evaluation method is as follows:

[0037] The yield volatility index α for each eligible wealth management product corresponding to each bank customer within the target bank during a preset time period. ir d1, Yield Growth Index α ir d2, Net Asset Value Growth Index α ir d3 and the information of the managers of each eligible financial product conform to the weighted index β. ir Substitute into the formula Obtain the recommendation coefficient Ψ for each eligible wealth management product for each bank customer within the target bank. ir , where τ1, τ2, and τ3 represent the preset proportional factors affecting the rate of return of the wealth management product, the proportional factors affecting the net asset value per unit, and the proportional factors affecting the management personnel information, respectively, and e represents the natural constant.

[0038] Compared with existing technologies, the product recommendation method based on artificial intelligence described in this invention has the following advantages:

[0039] This invention acquires asset information of customers within a target bank, analyzes the compliance ratio of their corresponding account information and product holdings with various risk types, further evaluates the risk preference prediction coefficients for each risk type, and compares these to obtain the corresponding risk types for each customer. This enables data-driven, systematic, and intelligent analysis of customer risk types, effectively avoiding the influence of individual psychological factors. This improves the certainty and accuracy of the risk type analysis results, ensuring that the financial products recommended by the bank's online platform match the customer's actual risk tolerance. This significantly enhances the accuracy of financial product recommendations on the bank's online platform and meets the financial investment needs of bank customers.

[0040] This invention extracts relevant data from each bank's customers within a target bank regarding eligible wealth management products, evaluates the recommendation coefficient of each product for each customer, and sorts the products in descending order of recommendation coefficient. This data is then pushed to the corresponding bank customers, enabling them to quickly find wealth management products that meet their expectations. This effectively avoids the problem of customers spending a lot of time comparing products one by one, further improving the customer's purchasing experience, increasing their enthusiasm for purchasing wealth management products, and ultimately boosting the sales volume of wealth management products on the bank's online platform, thus promoting the rapid development of the bank's online platform. Attached Figure Description

[0041] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the description of the embodiments will be briefly introduced below. 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 effort.

[0042] Figure 1 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0043] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0044] Please see Figure 1As shown, the present invention provides a product recommendation method based on artificial intelligence, comprising the following steps:

[0045] S1. Obtaining Bank Customer Asset Information: Obtaining asset information of customers of each bank within the target bank, including account information and product holding information.

[0046] Based on the above embodiments, the account information includes age, gender, average monthly income and average monthly expenditure, and the product holding information includes the risk type, holding period, holding amount and rate of return of each product held.

[0047] It should be noted that the risk types of the products held include, but are not limited to: conservative, moderate, and aggressive.

[0048] S2. Bank Customer Asset Information Analysis: Analyze the asset information of each bank customer within the target bank to obtain the compliance ratio index of each bank customer's corresponding account information and held product information with each risk type.

[0049] Based on the above embodiments, the method for parsing the compliance ratio index of each risk type for the account information corresponding to each bank customer within the target bank in step S2 is as follows:

[0050] Based on the account information of customers in each bank within the target bank, extract the age, gender, average monthly income, and average monthly expenditure of customers in each bank within the target bank. Subtract the average monthly expenditure of the corresponding bank customer from the average monthly income of customers in each bank within the target bank to obtain the average monthly savings amount of customers in each bank within the target bank.

[0051] Extract the risk tolerance weights for each risk type corresponding to customers of different age groups, genders, and average monthly savings amounts stored in the bank's data repository. Based on the age, gender, and average monthly savings amount of customers within the target bank, filter the risk tolerance weights for each risk type corresponding to the age, gender, and average monthly savings amount of customers within the target bank, and label them as p. i c a1、p i c a2、p i c a3, i = 1, 2, ..., n, where i represents the ID of the i-th bank customer, and c = 1, 2, ..., k, where c represents the ID of the c-th risk type;

[0052] Analysis of the compliance ratio index of customer account information for each risk type within the target bank. δ1, δ2, and δ3 represent the risk type compliance factors corresponding to the preset customer age, customer gender, and average monthly savings amount, respectively.

[0053] Based on the above embodiments, the method for analyzing the compliance ratio index of the product information held by each bank customer within the target bank for each risk type in step S2 is as follows:

[0054] Based on the product holding information of customers within the target bank, extract the risk type, holding period, holding amount, and rate of return for each product held by each customer within the target bank. Count the number of products held by each customer within the target bank for each risk type and label them as x. i c ;

[0055] Extract the holding time, holding amount, and rate of return of each product held by each bank customer within the target bank for each risk type, and obtain the preference of each bank customer for each risk type of held products within the target bank, which is denoted as φ. i c ;

[0056] Analysis of the compliance ratio index of product holdings of customers of each bank within the target bank for each risk type Where μ represents the preset correction factor for customer-held product information, and e represents the natural constant.

[0057] Furthermore, the analytical formula for analyzing the preference of customers of each bank within the target bank for products held by different risk types is as follows: Where φ i c p represents the preference of the i-th bank customer within the target bank for holding products corresponding to the c-th risk type, where γ1, γ2, and γ3 represent the preference influencing factors corresponding to the preset product holding time, product holding amount, and product yield, respectively. i c b j 1. p i c b j 2. p i c b j 3 represents the holding time, holding amount, and rate of return of the j-th product held by the i-th bank customer in the c-th risk type within the target bank, where j = 1, 2, ..., m, and k represents the number of risk types.

[0058] S3. Risk Preference Prediction Coefficient Assessment: Based on the compliance ratio index of each bank customer's account information and product holding information for each risk type within the target bank, assess the risk preference prediction coefficient for each risk type within the target bank.

[0059] Based on the above embodiments, the method for assessing the risk preference prediction coefficients for each risk type of each bank customer within the target bank in step S3 is as follows:

[0060] The compliance ratio index ξ of customer account information of each bank within the target bank for each risk type. i c The compliance ratio index ψ of the product holding information of each bank's customers with each risk type i c Substitute into the risk preference prediction coefficient assessment formula θ i c =λ1*ξ i c +λ2*ψ i c The risk preference prediction coefficients θ for each risk type of customers within the target bank are obtained. i c λ1 and λ2 represent the weighting factors corresponding to the preset customer account information and the weighting factors corresponding to the customer's product information, respectively, and λ1+λ2=1.

[0061] S4. Selection of Eligible Wealth Management Products: Compare the risk preference prediction coefficients of customers of each bank within the target bank for each risk type to obtain the risk type corresponding to customers of each bank within the target bank, and select eligible wealth management products for customers of each bank within the target bank.

[0062] Based on the above embodiments, the specific steps corresponding to step S4 are as follows:

[0063] Compare the risk preference prediction coefficients of customers of each bank in the target bank for each risk type, and select the risk type with the highest risk preference prediction coefficient for customers of each bank in the target bank, and record it as the risk type for customers of each bank in the target bank.

[0064] Extract the risk types corresponding to each wealth management product from the target bank's platform. Based on the risk types corresponding to each bank's customers within the target bank, filter out the wealth management products whose risk types match those of each bank's customers within the target bank, and record them as the eligible wealth management products corresponding to each bank's customers within the target bank.

[0065] In this embodiment, the present invention obtains asset information of customers of each bank within the target bank, analyzes the conformity ratio index of the corresponding account information and product holding information of each customer within the target bank to each risk type, further evaluates the risk preference prediction coefficient of each customer within the target bank for each risk type, and compares them to obtain the risk type corresponding to each customer within the target bank. This enables data-driven, systematic, and intelligent analysis of the risk type of bank customers, effectively avoiding the influence of personal psychological factors, thereby improving the certainty and accuracy of the risk type analysis results of bank customers. This ensures that the financial products recommended by the bank's online platform to users are in line with the customer's actual risk tolerance, greatly improving the accuracy of financial product recommendations on the bank's online platform and meeting the financial investment needs of bank customers.

[0066] S5. Data Extraction for Eligible Wealth Management Products: Extract relevant data for each eligible wealth management product for each bank customer within the target bank. The relevant data includes yield curves, unit net value curves, and management personnel information.

[0067] Based on the above embodiments, step S5 extracts relevant data for each eligible wealth management product corresponding to each bank customer within the target bank, specifically including:

[0068] Extract relevant data of each bank's customers corresponding to each eligible wealth management product from the target bank's platform, and obtain the yield curve, unit net value curve and management personnel information of each bank's customers corresponding to each eligible wealth management product within the target bank.

[0069] Based on the yield curves and net asset value curves of each eligible wealth management product for each bank's customers within the target bank, the yields and net asset values ​​of each eligible wealth management product for each bank's customers within the target bank at each data collection point within a preset time period are obtained. The yield volatility index, yield growth index, and net asset value growth index of each eligible wealth management product for each bank's customers within the target bank during the preset time period are analyzed and denoted as α. ir d1, α ir d2, α ir d3, r = 1, 2, ..., u, where r represents the r-th product number that meets the criteria for a financial product;

[0070] Based on the information of the management personnel of each eligible wealth management product for each customer within the target bank, a weighted index for the management personnel information of each eligible wealth management product for each customer within the target bank is obtained and labeled as β. ir .

[0071] Based on the above embodiments, the formula for analyzing the yield volatility index of each eligible wealth management product for each bank customer within the target bank during a preset time period is as follows: Where αir d1 represents the yield volatility index of the r-th eligible wealth management product for the i-th bank customer within the target bank during the preset time period, h represents the number of data collection points within the preset time period, and y ir d f Let f represent the yield of the r-th eligible wealth management product for the i-th bank customer within the target bank at the f-th data collection point within the preset time period, where f = 1, 2, ..., h. This represents the allowable standard deviation corresponding to the preset yield of the wealth management product.

[0072] Furthermore, the target bank's customers correspond to the yield growth index of each eligible wealth management product within a preset time period. Where α ir d2 represents the yield growth index of the r-th eligible wealth management product for the i-th bank customer within the target bank during the preset time period, η represents the preset yield growth compensation factor for the wealth management product, and y ir d h Let y represent the yield of the r-th eligible wealth management product for the i-th bank customer within the target bank at the h-th data collection point within the preset time period. ir d1 represents the yield of the r-th eligible wealth management product for the i-th bank customer within the target bank at the first data collection point within the preset time period.

[0073] Furthermore, the analysis formula for the unit net value growth index of each bank customer within the target bank corresponding to each eligible wealth management product within a preset time period is as follows: Where α ir d3 represents the unit net asset value growth index of the r-th eligible wealth management product for the i-th bank customer within the target bank during the preset time period, χ represents the preset unit net asset value growth correction factor for the wealth management product, and z ir d f+1 Let z represent the unit net value of the r-th eligible wealth management product for the i-th bank customer within the target bank at the (f+1)-th data collection point within the preset time period. ir d f This represents the unit net value of the r-th eligible wealth management product for the i-th bank customer within the target bank at the f-th data collection point within the preset time period.

[0074] Based on the above embodiments, the weighted index parsing method for the management personnel information of each bank customer within the target bank corresponding to each eligible wealth management product is as follows:

[0075] Extract the management personnel information for each eligible wealth management product corresponding to each bank customer within the target bank. This information includes the manager's years of wealth management experience, average annualized return, and cumulative managed amount. Label the years of wealth management experience, average annualized return, and cumulative managed amount in the management personnel information for each eligible wealth management product corresponding to each bank customer within the target bank as w. ir g1, w ir g2、w ir g3,

[0076] Extract the average years of wealth management experience, average annualized return, and average cumulative managed amount of wealth management product managers from the target bank platform, and label them as follows:

[0077] Based on the management personnel information, it conforms to the weighted index analysis formula. Information on the management personnel of each eligible wealth management product for each bank's customers within the target bank is obtained, conforming to the weighted index β. ir ε1, ε2, and ε3 represent the default influencing factors corresponding to the management personnel's years of financial management experience, average annualized rate of practice, and cumulative amount of management, respectively.

[0078] S6. Evaluation of Recommendation Coefficient for Eligible Wealth Management Products: The evaluation assesses the recommendation coefficient of each eligible wealth management product for each bank customer within the target bank, and sorts them in descending order of recommendation coefficient, while simultaneously pushing the results to the corresponding bank customers.

[0079] Based on the above embodiments, in step S6, the recommendation coefficients for each eligible wealth management product for each bank customer within the target bank are evaluated. The specific evaluation method is as follows:

[0080] The yield volatility index α for each eligible wealth management product corresponding to each bank customer within the target bank during a preset time period. ir d1, Yield Growth Index α ir d2, Net Asset Value Growth Index α ir d3 and the information of the managers of each eligible financial product conform to the weighted index β. ir Substitute into the formula Obtain the recommendation coefficient Ψ for each eligible wealth management product for each bank customer within the target bank. ir , where τ1, τ2, and τ3 represent the preset proportional factors affecting the rate of return of the wealth management product, the proportional factors affecting the net asset value per unit, and the proportional factors affecting the management personnel information, respectively, and e represents the natural constant.

[0081] Furthermore, the recommendation coefficients of each bank customer within the target bank for each eligible wealth management product are compared with each other, and the products are sorted in descending order of recommendation coefficient. The top five eligible wealth management products for each bank customer within the target bank are extracted and recorded as the recommended wealth management products for each bank customer within the target bank, and then pushed to the corresponding bank customers.

[0082] In this embodiment, the present invention extracts relevant data on each eligible wealth management product for each bank customer within the target bank, evaluates the recommendation coefficient of each eligible wealth management product for each bank customer within the target bank, and sorts them in descending order of recommendation coefficient. This data is then pushed to the corresponding bank customers, enabling them to quickly find wealth management products that meet their expectations. This effectively avoids the problem of bank customers spending a lot of time comparing and reviewing products one by one, further improving the customer's purchasing experience, increasing their enthusiasm for purchasing wealth management products, and ultimately increasing the sales volume of wealth management products on the bank's online platform, thus promoting the rapid development of the bank's online platform.

[0083] The above description is merely an example and illustration of the concept of the present invention. Those skilled in the art can make various modifications or additions to the specific embodiments described or use similar methods to replace them, as long as they do not deviate from the concept of the invention or exceed the scope defined in the claims, they should all fall within the protection scope of the present invention.

Claims

1.A product push method based on artificial intelligence, characterized by, Comprise the following steps: S1, bank customer asset information acquisition: acquire the asset information of each bank customer in the target bank, wherein the asset information comprises account information and holding product information; S2, bank customer asset information analysis: analyze the asset information of each bank customer in the target bank to obtain the compliance proportion index of the corresponding account information and holding product information of each bank customer in the target bank to each risk type; S3, risk preference estimation coefficient evaluation: according to the compliance proportion index of the corresponding account information and holding product information of each bank customer in the target bank to each risk type, evaluate the risk preference estimation coefficient of each bank customer in the target bank to each risk type; S4, compliance financial product screening: compare the risk preference estimation coefficient of each bank customer in the target bank to each risk type, obtain the risk type corresponding to each bank customer in the target bank, and screen each compliance financial product corresponding to each bank customer in the target bank; S5, compliance financial product data extraction: extract the related data of each compliance financial product corresponding to each bank customer in the target bank, wherein the related data comprises yield curve, unit net value curve and manager information; S6, compliance financial product recommendation coefficient evaluation: evaluate the recommendation coefficient of each compliance financial product corresponding to each bank customer in the target bank, and sort them in order from high to low, and push to the corresponding bank customer at the same time; The holding product information comprises the risk type, holding time, holding amount and yield of each holding product; The compliance proportion index of the holding product information of each bank customer in the target bank to each risk type in step S2 is analyzed as follows: According to the holding product information of each bank customer in the target bank, the risk type, holding time, holding amount and yield of each holding product of each bank customer in the target bank are extracted, the number of holding products of each risk type corresponding to each bank customer in the target bank is counted, and the number of holding products of each risk type corresponding to each bank customer in the target bank is marked as ; Extract the holding time, holding amount and yield of each holding product of each risk type corresponding to each bank customer in the target bank to obtain the preference degree of each bank customer in the target bank for each holding product of each risk type, and mark it as ; The analysis target bank is parsed to each bank customer corresponding holding product information to each risk type of the coincidence proportion index , wherein is expressed as a preset customer holding product information coincidence correction factor, e is expressed as a natural constant, , i is expressed as the number of the i-th bank customer, , and c is expressed as the number of the c-th risk type. 2.The product pushing method based on artificial intelligence according to claim 1, characterized in that: The compliance proportion index of the account information of each bank customer in the target bank to each risk type in step S2 is analyzed as follows: According to the account information of each bank customer in the target bank, the age, gender, average monthly income and average monthly expenditure of each bank customer in the target bank are extracted, the average monthly income of each bank customer in the target bank is subtracted by the average monthly expenditure of the corresponding bank customer, and the average monthly savings of each bank customer in the target bank is obtained; The set risk bearing capacity weights of each risk type corresponding to the customers of each age group, the customers of each gender, and the average monthly savings amount ranges stored in the bank data repository are extracted, and the set risk bearing capacity weights of each risk type corresponding to the age, gender, and average monthly savings amount of each bank customer in the target bank are screened according to the age, gender, and average monthly savings amount of each bank customer in the target bank, and are marked as ; The risk type compliance proportion index of each bank customer corresponding account information in the target bank is analyzed wherein respectively represent the risk type compliance influence factor corresponding to the preset customer age, customer gender, and customer average monthly savings amount. 3.The product pushing method based on artificial intelligence according to claim 1, characterized in that: The risk preference estimation coefficient of each bank customer in the target bank to each risk type in step S3 is evaluated as follows: a corresponding account information of each bank customer in the target bank to each risk type a corresponding holding product information of each bank customer to each risk type a risk preference estimation coefficient evaluation formula a risk preference estimation coefficient of each bank customer in the target bank to each risk type wherein respectively represent a preset corresponding compliance weight factor of the customer account information and a corresponding compliance weight factor of the customer holding product information . 4.The product pushing method based on artificial intelligence according to claim 3, characterized in that: The specific steps in step S4 are as follows: Compare the risk preference estimation coefficient of each bank customer in the target bank to each risk type, screen the risk type with the highest risk preference estimation coefficient corresponding to each bank customer in the target bank, and mark it as the risk type corresponding to each bank customer in the target bank; Extract the risk type corresponding to each financial product from the target bank platform, according to the risk type corresponding to each bank customer in the target bank, screen each financial product corresponding to the risk type of each bank customer in the target bank, and mark it as each compliance financial product corresponding to each bank customer in the target bank. 5.The product pushing method based on artificial intelligence according to claim 1, characterized in that: In step S5, the related data of each compliance financial product corresponding to each bank customer in the target bank is extracted, which specifically comprises: Extracting the related data of each bank customer corresponding to each qualified wealth management product from the target bank platform, obtaining the yield curve, unit net value curve and manager information of each bank customer corresponding to each qualified wealth management product in the target bank; According to the yield curve and the unit net value curve of each financial product corresponding to each bank customer in the target bank, the yield and the unit net value of each financial product corresponding to each bank customer in the target bank at each collection time point in the preset time period are obtained, and the yield fluctuation index, the yield growth index and the unit net value growth index of each financial product corresponding to each bank customer in the target bank in the preset time period are analyzed, which are respectively marked as , , , , r represents the number of the rth financial product; According to the information of the managers corresponding to each financial product that meets the requirements of each bank customer in the target bank, the information of the managers corresponding to each financial product that meets the requirements of each bank customer in the target bank is analyzed to obtain the weight index, which is marked as . 6.The product pushing method based on artificial intelligence according to claim 5, characterized in that: The yield rate fluctuation index analysis formula of each bank product corresponding to each bank customer in the target bank in a preset time period is , wherein represents the yield rate fluctuation index of the rth bank product corresponding to the ith bank customer in the target bank in a preset time period, h represents the number of collection time points in the preset time period, represents the yield rate of the rth bank product corresponding to the ith bank customer in the target bank at the fth collection time point in the preset time period, , represents the preset allowed standard deviation value corresponding to the yield rate of the bank product. 7.The product pushing method based on artificial intelligence according to claim 5, characterized in that: The manager information of each bank customer corresponding to each qualified wealth management product in the target bank conforms to a weight index analysis method, and the weight index analysis method is as follows: extracting the information of the managers corresponding to each of the financial products from each of the bank customers in the target bank, wherein the information of the managers includes the management years of the managers, the average annual rate of employment, and the cumulative management amount, marking the management years of the managers, the average annual rate of employment, and the cumulative management amount in the information of the managers corresponding to each of the financial products from each of the bank customers in the target bank as , respectively, Extract the average wealth management years, average annual rate of employment, and average cumulative management amount of the wealth management product managers corresponding to the target bank platform, and mark them as ; According to the management personnel information meets the weight index analytical formula , the target bank customer corresponding to each financial product management personnel information meets the weight index is obtained , wherein respectively represent the preset management personnel corresponding to the financial management years, the average annual rate of employment, the cumulative management amount corresponding to the compliance influence factor. 8.The product pushing method based on artificial intelligence according to claim 5, characterized in that: The recommendation coefficient of each bank customer corresponding to each qualified wealth management product in the target bank is evaluated in the step S6, and the specific evaluation method is as follows: The yield rate fluctuation index of each bank customer in the target bank corresponding to each qualified financial product in a preset time period , the yield rate growth index , the unit net value growth index , and the manager information of each qualified financial product Substitute the formula , and obtain the recommendation coefficient of each bank customer in the target bank corresponding to each qualified financial product , wherein , respectively, are the preset yield rate influence proportion factor, the unit net value influence proportion factor, and the manager information influence proportion factor corresponding to the financial product, and e is the natural constant.

Images