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

Abstract

The embodiment of the invention belongs to the field of financial science and technology, and relates to a financial product recommendation method and device, computer equipment and a storage medium, and the method comprises the following steps: obtaining financial real-time data and a user portrait; constructing a risk preference model based on the user portrait, and obtaining a user preference product combination according to the risk preference model; obtaining an expected return rate, and screening the user preference product combination according to the expected return rate to obtain a recommended product combination; training a preset income prediction model according to the recommended product combination to obtain an income risk balance model; constructing a purchase preference model based on the user portrait, and sorting the user preference product combinations according to the purchase preference model and the income risk balance model to obtain a product combination sorting table; and product combination pushing is carried out on the user according to the product combination sorting table. According to the invention, personalized and accurate financial product combination recommendation can be effectively realized for the user.

Description

Technical Field

[0001] The present application relates to the field of data processing technology, specifically to the field of financial technology, and in particular to a method, device, computer equipment and storage medium for recommending financial products. Background Art

[0002] In the current financial market, personal financial advice is usually provided manually by financial advisors based on factors such as the individual's financial situation, risk tolerance and investment horizon. However, this traditional approach usually has several problems: first, it relies on the subjective judgment of individual advisors, second, it is inefficient when handling a large number of customer inquiries, and third, it is unable to update investment advice in real time to adapt to market changes. In addition, most existing automated recommendation systems focus mainly on stocks or bonds, rather than a wider portfolio of products, including conservative investment products such as time deposits. Bank deposit products on the current market usually offer fixed interest rates, leaving customers with limited returns. Although financial products may offer higher returns, these product portfolios are usually accompanied by higher risks, and the degree of risk is often unknown to customers.

[0003] Due to lack of professional knowledge and tools, many customers can only rely on their own judgment and preferences to choose product combinations, which may lead to low returns or even losses. In addition, bank account managers can only rely on individual communication with customers and the characteristics of the product combination itself when recommending product combinations. This method is not only inefficient, but also lacks personalization and objectivity, may not be suitable for all customers, and there is a lag in product combination information, which leads to reduced accuracy of product recommendations and customer satisfaction. Summary of the invention

[0004] The purpose of the embodiments of the present application is to provide a financial product recommendation method, device, computer equipment and storage medium to solve the problem of being unable to accurately and effectively push financial product combinations.

[0005] In order to solve the above technical problems, the present application embodiment provides a financial product recommendation method, which adopts the following technical solution:

[0006] Obtain real-time financial data and user portraits;

[0007] Building a risk preference model based on the user portrait, and obtaining the user's preferred product portfolio according to the risk preference model;

[0008] Obtaining an expected rate of return, and screening the user's preferred product portfolio according to the expected rate of return to obtain a recommended product portfolio;

[0009] Training a preset revenue prediction model according to the recommended product portfolio to obtain a revenue-risk balance model;

[0010] Building a purchase preference model based on the user portrait, and sorting the user's preferred product portfolio according to the purchase preference model and the benefit-risk balance model to obtain a product portfolio sorting table;

[0011] Product combinations are pushed to users according to the product combination ranking table.

[0012] Furthermore, the step of obtaining real-time financial data and user portraits specifically includes:

[0013] Extracting the financial real-time data from a financial database in real time;

[0014] Obtaining user collected information, and organizing and analyzing the user collected information to obtain user key characteristics and user trends;

[0015] The user portrait is constructed according to the user key features and the user trends.

[0016] Furthermore, the step of building a risk preference model based on the user portrait and obtaining the user's preferred product portfolio according to the risk preference model specifically includes:

[0017] Acquire the user's historical purchase data and psychological evaluation data based on the user portrait;

[0018] Constructing the risk preference model based on the user's historical purchase data and the psychological assessment data;

[0019] Products that meet the user's risk level are screened and combined according to the risk preference model to obtain the user's preferred product combination.

[0020] Furthermore, the step of obtaining the expected rate of return, screening the user's preferred product combination according to the expected rate of return, and obtaining the recommended product combination specifically includes:

[0021] Obtain user account information and market product data;

[0022] Acquire the expected rate of return according to the user account information;

[0023] The user preferred product combination is screened according to the expected rate of return and the market product data to obtain the recommended product combination.

[0024] Furthermore, the step of training a preset revenue prediction model according to the recommended product portfolio to obtain a revenue-risk balance model specifically includes:

[0025] Inputting the recommended product combination into the preset profit prediction model to perform profit calculation to obtain the product combination rate of return;

[0026] Obtaining a preset expected rate of return, and screening the recommended product combination according to the preset expected rate of return and the product combination rate of return to obtain an expected recommended product combination;

[0027] Calculating the optimal solution for the expected recommended product combination based on a simulated annealing algorithm to obtain an optimal recommended product combination;

[0028] The optimal recommended product combination is verified, and after the verification is passed, the preset profit prediction model and the simulated annealing algorithm are combined and output as the profit-risk balance model.

[0029] Furthermore, the step of constructing a purchase preference model based on the user portrait, sorting the user preferred product combinations according to the purchase preference model and the benefit-risk balance model, and obtaining a product combination sorting table specifically includes:

[0030] Acquire the user's historical purchase data according to the user portrait;

[0031] Parsing the user's historical purchase data to obtain the user's historically purchased products and the user's purchase amount;

[0032] Constructing the purchase preference model based on the user's historical purchases of products and the amount of money the user spent on the products;

[0033] Inputting the user's preferred product combination into the purchase preference model to obtain a first model score;

[0034] Inputting the user's preferred product portfolio into the return-risk balance model to obtain a second model score;

[0035] Calculate a comprehensive model score according to the first model score and the second model score;

[0036] The user preferred product combinations are sorted according to the comprehensive scores of the model to obtain a product combination sorting table.

[0037] Furthermore, after the step of pushing product combinations to users according to the product combination ranking table, the following steps are also included:

[0038] Collect product push feedback data;

[0039] Calculating a deviation value between the product push feedback data and the expected feedback data of the product combination ranking table;

[0040] Determine whether the deviation value is greater than or equal to a preset threshold;

[0041] If the deviation value is greater than or equal to the preset threshold, the purchase preference model and the return-risk balance model are evaluated according to the deviation value, and the purchase preference model and the return-risk balance model are optimized according to the evaluation results to obtain the optimized purchase preference model and the optimized return-risk balance model, and a new product combination ranking table is regenerated to push product combinations to users until the deviation value corresponding to the new product combination ranking table is less than the preset threshold.

[0042] In order to solve the above technical problems, the embodiment of the present application further provides a financial product recommendation device, which adopts the following technical solution:

[0043] Data acquisition module, used to obtain real-time financial data and user portraits;

[0044] A product portfolio acquisition module, used to build a risk preference model based on the user portrait, and acquire the user's preferred product portfolio according to the risk preference model;

[0045] A product combination screening module is used to obtain an expected rate of return, and screen the user's preferred product combination according to the expected rate of return to obtain a recommended product combination;

[0046] A model training module is used to train a preset revenue prediction model according to the recommended product portfolio to obtain a revenue-risk balance model;

[0047] A product combination ranking module, used to build a purchase preference model based on the user portrait, and to sort the user preferred product combinations according to the purchase preference model and the benefit-risk balance model to obtain a product combination ranking table;

[0048] The product combination push module is used to push product combinations to users according to the product combination ranking table.

[0049] In order to solve the above technical problems, the embodiment of the present application further provides a computer device, which adopts the following technical solution:

[0050] A computer device comprises a memory and a processor, wherein the memory stores computer-readable instructions, and the processor implements the steps of any one of the above financial product recommendation methods when executing the computer-readable instructions.

[0051] In order to solve the above technical problems, the embodiment of the present application further provides a computer-readable storage medium, which adopts the following technical solution:

[0052] A computer-readable storage medium having computer-readable instructions stored thereon, wherein the computer-readable instructions, when executed by a processor, implement the steps of the financial product recommendation method as described in any one of the above.

[0053] Compared with the prior art, the embodiments of the present application have the following beneficial effects: the embodiments obtain real-time financial data and user portraits; build a risk preference model based on the user portraits, and obtain the user's preferred product portfolio according to the risk preference model; obtain the expected rate of return, screen the user's preferred product portfolio according to the expected rate of return, and obtain the recommended product portfolio; train the preset income prediction model according to the recommended product portfolio to obtain the income-risk balance model; build a purchase preference model based on the user portraits, sort the user's preferred product portfolio according to the purchase preference model and the income-risk balance model, and obtain a product portfolio ranking table; push product portfolios to users according to the product portfolio ranking table. This effectively realizes personalized and accurate financial product portfolio recommendations for users, so as to provide users with accurate and effective investment advice and references. BRIEF DESCRIPTION OF THE DRAWINGS

[0054] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0055] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0056] Figure 2 A flowchart of an embodiment of a financial product recommendation method according to the present application;

[0057] Figure 3 yes Figure 2 A flowchart of a specific implementation of step S10;

[0058] Figure 4 yes Figure 2 A flowchart of a specific implementation of step S20;

[0059] Figure 5 yes Figure 2 A flowchart of a specific implementation of step S30;

[0060] Figure 6 yes Figure 2 A flowchart of a specific implementation of step S40;

[0061] Figure 7 yes Figure 2 A flowchart of a specific implementation of step S50;

[0062] Figure 8 is a structural schematic diagram of an embodiment of a financial product recommendation device according to the present application;

[0063] Fig. 9 It is a structural diagram of an embodiment of a computer device according to the present application. DETAILED DESCRIPTION

[0064] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0065] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor does it refer to non-related or alternative embodiments that are mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0066] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0067] like Figure 1 As shown, the system architecture 100 may include terminal devices 101, 102, 103, a network 104 and a server 105. The network 104 is used to provide a medium for communication links between the terminal devices 101, 102, 103 and the server 105. The network 104 may include various connection types, such as wired, wireless communication links or optical fiber cables, etc.

[0068] Users can use terminal devices 101, 102, 103 to interact with server 105 through network 104 to receive or send messages, etc. Various communication client applications can be installed on terminal devices 101, 102, 103, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0069] Terminal devices 101, 102, 103 can be various electronic devices with display screens and supporting web browsing, including but not limited to smart phones, tablet computers, e-book readers, MP3 players (Moving Picture Experts Group Audio Layer III), MP4 (Moving Picture Experts Group Audio Layer IV) players, laptop computers, desktop computers, etc.

[0070] The server 105 may be a server that provides various services, such as a background server that provides support for web pages displayed on the terminal devices 101 , 102 , and 103 .

[0071] It should be noted that the financial product recommendation method provided in the embodiment of the present application is generally executed by a server, and accordingly, the financial product recommendation device is generally arranged in the server.

[0072] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0073] Continue to refer Figure 2 , shows a flow chart of an embodiment of a method for recommending financial products according to the present application. The financial product recommendation method comprises the following steps:

[0074] Step S10, obtaining financial real-time data and user portraits;

[0075] In this embodiment, financial real-time data includes macroeconomic indicators, bank deposit rates, stock market indices, product yield series, market product data, etc. User portraits are constructed based on user financial data, which include age, occupation, income level, asset status, debt status, historical investment preferences, user historical purchase data, psychological assessment data, etc.

[0076] Step S20, constructing a risk preference model based on the user portrait, and obtaining the user's preferred product portfolio according to the risk preference model;

[0077] In this embodiment, the risk preference model is a model used to determine the risk level that the user can bear and to obtain the corresponding product portfolio based on the risk level. The product portfolio refers to a combination of financial business products, specifically a combination of two or more of investment business products, wealth management business products, deposit business products, and fund business products. The risk preference model can effectively obtain the financial business product portfolio preferred by the user under the current risk level.

[0078] Step S30, obtaining an expected rate of return, and screening the user's preferred product portfolio according to the expected rate of return to obtain a recommended product portfolio;

[0079] In this embodiment, the expected rate of return corresponds to the user's expected return. The expected rate of return can be obtained by conducting a survey on the user or in the form of a questionnaire when the user first logs in to the business system. By screening the user's preferred products based on the expected rate of return, the recommended financial business product portfolio under the current expected rate of return can be effectively obtained.

[0080] Step S40, training a preset revenue prediction model according to the recommended product combination to obtain a revenue-risk balance model;

[0081] In this embodiment, the preset profit prediction model is trained by using historical sample market data and existing sample market data to obtain a profit prediction model that can effectively predict the profit of the recommended product combination under the current market conditions. By training the profit prediction model with the recommended product combination, a profit-risk balance model based on profit and risk considerations is obtained.

[0082] Step S50, constructing a purchase preference model based on the user portrait, and sorting the user's preferred product combinations according to the purchase preference model and the benefit-risk balance model to obtain a product combination sorting table;

[0083] In this embodiment, the purchase preference model is a model constructed by considering the user's purchasing behavior. By constructing the purchase preference model, the user's purchasing preference can be effectively grasped to achieve personalized and accurate product combination push. The optimal product combination plan is pushed to the user through the benefit-risk balance model to achieve effective product combination push. The purchase preference model and the benefit-risk balance model are combined to rank the user's preferred product combinations, so that the resulting product combination ranking table not only conforms to the user's personality, but also improves accuracy and reliability.

[0084] Step S60: Push product combinations to users according to the product combination ranking table.

[0085] In this embodiment, the product combination ranking table is ranked from high to low according to the comprehensive total scores of the models evaluated by the purchase preference model and the risk-return balance model. The comprehensive total scores of the models reflect the priority recommended by the user's preferred product combinations in the product combination ranking table. When the product combination is pushed to the user, the user's preferred products in the product combination ranking table are sent in order to the user's business display interface for interface rendering and display, thereby realizing the step of pushing product combinations to the user according to the product combination ranking table.

[0086] This embodiment obtains real-time financial data and user portraits; builds a risk preference model based on the user portrait, and obtains the user's preferred product portfolio according to the risk preference model; obtains the expected rate of return, screens the user's preferred product portfolio according to the expected rate of return, and obtains a recommended product portfolio; trains a preset return prediction model according to the recommended product portfolio to obtain a return-risk balance model; builds a purchase preference model based on the user portrait, sorts the user's preferred product portfolio according to the purchase preference model and the return-risk balance model, and obtains a product portfolio ranking table; pushes product portfolios to users according to the product portfolio ranking table. This effectively realizes personalized and accurate financial product portfolio recommendations for users, so as to provide users with accurate and effective investment advice and references.

[0087] Continue to refer Figure 3 In some optional implementations of this embodiment, step S10 includes the following steps:

[0088] Step S101, extracting the financial real-time data from a financial database in real time;

[0089] In this embodiment, financial real-time data is extracted from the financial database through financial data extraction identifiers. The financial real-time data includes macroeconomic indicators, bank deposit interest rates, stock market indices, product yield series, market product data, etc. Among them, macroeconomic indicators can be obtained and updated through statistical data of government departments, research and analysis of financial institutions and research institutions, bank deposit interest rates can be obtained and updated through policy information published by the central bank and commercial banks, stock market indices can be obtained and updated through real-time stock market index data provided by securities exchanges, and financial data provided by financial data service providers, product yield series can be obtained and updated through product yield information announced by financial product issuers, and financial product yield data collected, organized and analyzed by data service providers, and market product data can be obtained and updated through product data information collected and published by exchanges, financial institutions, and third-party data service providers.

[0090] Step S102, obtaining user collected information, and sorting and analyzing the user collected information to obtain user key features and user trends;

[0091] In this embodiment, user collected information is obtained through multiple channels, including personal information filled in by the user when registering (such as name, age, occupation, contact information, etc.), user transaction behavior data (such as purchase records of financial products, transaction amount, transaction time, etc.), user feedback data (such as questionnaire results, customer service consultation records, etc.) and possible external data (such as social media data, credit records, etc.). Before collating and analyzing the user collected information, it is necessary to perform data cleaning on the user collected information to remove duplicate, erroneous, incomplete or irrelevant data. Data cleaning includes operations such as format unification, missing value processing, outlier detection and correction, etc. By performing data cleaning operations such as format unification, missing value processing, outlier detection and correction on the user collected information, complete and effective user collected information can be obtained. The collation and analysis of user collected information includes data collation and data analysis. Data collation is to organize the cleaned user collected information according to a certain logic and structure, such as classification, sorting, grouping, etc., to obtain orderly user collected information. Data analysis is to use statistical analysis and data mining techniques to conduct in-depth analysis of user collected information to mine the key characteristics and trends of users, including user behavior pattern analysis, user preference analysis, user risk tolerance assessment, etc. By conducting user behavior pattern analysis, user preference analysis, and user risk tolerance assessment on user collected information, we can effectively obtain user key characteristics and user trends.

[0092] Step S103: construct the user portrait according to the user key features and the user trends.

[0093] In this embodiment, the key features of the user include age, gender, income level, investment experience, risk preference, purchase preference, etc. The user trend is the trend change information of the corresponding key features of the user, such as income level changes from low to high, risk preference changes from high to low, etc. The construction of the user portrait is based on the extracted key features and trends of the user, and a comprehensive portrait of the user is constructed by using data visualization and labeling technology. The portrait should include multiple dimensions such as the user's basic information, investment preferences, risk tolerance, potential needs, etc. In this embodiment, after the user portrait is constructed, it needs to be updated in real time to ensure its accuracy and timeliness. The update of the user portrait is achieved by re-evaluating the key features and trends of the user, and adjusting and optimizing the portrait according to the new data.

[0094] This embodiment extracts the financial real-time data from the financial database in real time; obtains user collection information, and organizes and analyzes the user collection information to obtain user key features and user trends; and constructs the user portrait according to the user key features and user trends. In this way, the financial real-time data related to financial business products and the user portrait related to the user are effectively obtained, providing effective and reliable data support for subsequent processing.

[0095] Continue to refer Figure 4 In some optional implementations of this embodiment, step S20 includes the following steps:

[0096] Step S201, obtaining user historical purchase data and psychological evaluation data according to the user portrait;

[0097] In this embodiment, the corresponding user financial data is obtained according to the user portrait, and the user financial data is searched to obtain the user's historical purchase data and psychological assessment data. The user's historical purchase data includes the type, price, purchase time, purchase frequency, return rate, evaluation information, etc. of the purchased products. The user's historical purchase data can be obtained from e-commerce platforms, CRM systems, payment records, etc. The psychological assessment data can be obtained by user access through professional psychological assessment questionnaires or tools, such as risk tolerance questionnaires, personality tests, etc. The assessment content of the above psychological assessment questionnaires or tools includes the user's risk attitude (such as conservative, balanced, aggressive), personality type (such as introversion / extroversion, intuition / feeling, thinking / emotion, judgment / perception), etc. After obtaining the above psychological assessment data, the assessment results are quantified and converted into analyzable values ​​or levels to obtain psychological assessment data.

[0098] Step S202, constructing the risk preference model based on the user's historical purchase data and the psychological assessment data;

[0099] In this embodiment, key features such as purchase frequency, average consumption, product type distribution, risk attitude level, etc. are extracted by feature engineering combined with user historical purchase data and psychological assessment data, and then machine learning or statistical methods (such as decision tree, random forest, logistic regression, neural network, etc.) are used to build a model. In this embodiment, the risk preference model uses a decision tree model, and the risk preference model is constructed based on the formula R=f(H, r, rl_i), where H is the user's historical purchase data product, which is used to adjust the customer risk level, r is the customer risk level, and rl_i is the product-specific risk level obtained from the product risk database. In specific implementation, by combining the features (H) of the user's historical purchase data, the customer risk level (r) and the product-specific risk level (rl_i) into a feature vector, using a decision tree algorithm, taking the feature vector as input, and the user's risk preference level (R) as the target variable for model training, the performance of the model is evaluated by methods such as cross-validation, such as accuracy, recall, etc., and the model parameters are adjusted according to the evaluation results, such as the depth of the tree, the number of leaf nodes, etc., to optimize the model performance, thereby obtaining an effective and reliable risk preference model.

[0100] Step S203, screening and combining products that meet the user's risk level according to the risk preference model to obtain the user's preferred product combination.

[0101] In this embodiment, the user's risk preference model is matched with the risk level of the product to screen out products that meet the user's risk level, and then product combinations are made based on factors such as the user's purchasing habits, budget constraints, and product diversity to obtain a user-preferred product combination.

[0102] This embodiment obtains the user's historical purchase data and psychological assessment data according to the user portrait; constructs the risk preference model based on the user's historical purchase data and the psychological assessment data; and screens and combines products that meet the user's risk level according to the risk preference model, thereby obtaining a user preference product combination that meets the risk that the user can bear, so as to facilitate subsequent screening operations based on expected rate of return.

[0103] Continue to refer Figure 5 In some optional implementations of this embodiment, step S30 includes the following steps:

[0104] Step S301, obtaining user account information and market product data;

[0105] In this embodiment, the user account information corresponds to the user's system account. The user account information includes the account name, account ID, account identification, etc., which can be automatically obtained through the system. The market product data includes detailed information on various financial products such as stocks, bonds, funds, insurance, and wealth management products, such as yield, risk level, investment period, etc., as well as the latest market data obtained through financial data service providers or exchanges, such as real-time stock prices, index trends, changes in bond yields, etc.

[0106] Step S302, obtaining the expected rate of return according to the user account information;

[0107] In this embodiment, the expected rate of return is the expected rate of return of the user, which can be obtained by asking the user for a survey or in the form of a questionnaire when the user logs in to the business system for the first time. In this embodiment, the user account information is associated with the expected rate of return, which can be obtained by extracting the associated information through the user account information. The expected rate of return can also be calculated by analyzing the user's past investment records and calculating the average rate of return or the risk-adjusted rate of return.

[0108] Step S303: Screen the user preferred product combination according to the expected rate of return and the market product data to obtain the recommended product combination.

[0109] In this embodiment, the screening of the user's preferred product portfolio includes the following steps: First, based on the product's price, rate of return, investment period and other attributes, screen out the product portfolio that meets the user's purchase conditions and expected rate of return requirements. Utilize modern portfolio theory (such as mean-variance model, risk parity model, etc.), combined with the user's specific needs and preferences, optimize the screened product portfolio to maximize returns while controlling risks. Backtest and risk evaluate the optimized product portfolio through historical data or Monte Carlo simulation and other methods to ensure that its performance meets expectations in different market environments. Finally, generate a recommended product portfolio while ensuring that the performance of the product portfolio meets expectations.

[0110] This embodiment obtains user account information and market product data; obtains the expected rate of return based on the user account information; and screens the user's preferred product portfolio based on the expected rate of return and the market product data, thereby obtaining a recommended product portfolio that meets the user's corresponding expected rate of return, so as to provide effective and reliable data support for subsequent model training.

[0111] Continue to refer Figure 6 In some optional implementations of this embodiment, step S40 includes the following steps:

[0112] Step S401, inputting the recommended product combination into the preset profit prediction model to perform profit calculation to obtain the product combination rate of return;

[0113] In this embodiment, the preset return prediction model is constructed based on a Bayesian network. Bayesian Networks (BN) is a graphical probability model used to represent the dependencies between variables. In return prediction, Bayesian networks can be used to model the interactions between various factors that affect returns (such as macroeconomic indicators, market sentiment, company fundamentals, etc.). The construction of the preset return prediction model includes the following steps: Network construction: First, based on historical data and expert knowledge, determine the key factors affecting the return rate and construct a dependency graph between them. Parameter learning: Use historical data to estimate the conditional probability distributions (CPDs) between nodes in the Bayesian network. This usually involves statistical inference techniques such as maximum likelihood estimation or Bayesian estimation. Reasoning: Given new market conditions (ie, evidence), use the Bayesian network for reasoning to calculate the posterior distribution of the return rate. The posterior distribution is calculated by using the Bayesian formula p(θ|x)=p(x|θ)p(θ) / p(x) (where θ represents the model parameters or variables and x is the observed data) to calculate the posterior distribution p(R|x,M_t). Here, p(x|θ) is the likelihood function, which represents the probability of observing data x under given parameters θ; p(θ) is the prior distribution; p(x) is the marginal likelihood function, which is usually used as a normalization constant. Since the calculation of the posterior distribution often involves high-dimensional integration, numerical methods (such as MCMC, variational Bayes, etc.) are needed to approximate the solution.

[0114] After the revenue prediction model is built, it is also necessary to cross-validate and evaluate the performance of the revenue prediction model. Cross-validation: divide the historical data into a training set and a test set, use the training set to train the model, and use the test set to evaluate the model's predictive performance. Cross-validation can further reduce the risk of overfitting. Performance evaluation: Use appropriate error evaluation methods (such as mean absolute error MAE, mean square error MSE, R^2 determination coefficient, etc.) to evaluate the model's predictive performance. When the error does not meet expectations, it is also necessary to adjust the structure or parameters of the revenue prediction model based on the error, so as to achieve adjustment and optimization of the revenue prediction model.

[0115] Step S402, obtaining a preset expected rate of return, screening the recommended product combination according to the preset expected rate of return and the product combination rate of return, and obtaining an expected recommended product combination;

[0116] In this embodiment, the preset expected rate of return is the expected rate of return set by the system for the recommended product portfolio, which can be adjusted according to actual conditions, or can be automatically set by the system according to factors such as the risk tolerance and investment objectives of the sample users. Screening is performed through the following steps: the calculated product portfolio rate of return is compared with the preset expected rate of return. If the product portfolio rate of return meets or exceeds the preset expected rate of return, it is regarded as the expected recommended product portfolio; otherwise, it may be necessary to readjust the recommended product portfolio or adjust the preset expected rate of return.

[0117] Step S403, calculating the optimal solution for the expected recommended product combination based on a simulated annealing algorithm to obtain an optimal recommended product combination;

[0118] In this embodiment, the simulated annealing algorithm is a heuristic random search algorithm used to find the global optimal solution in a large search space. The simulated annealing algorithm calculates the optimal solution of the desired recommended product combination including the following steps: Setting the objective function: The objective function is a balance function that maximizes the rate of return, minimizes the risk, or considers both the rate of return and the risk. Executing the algorithm: Through the simulated annealing algorithm, the product types, investment ratios and other parameters in the product combination are continuously adjusted to seek the optimal solution of the objective function. Obtaining the optimal recommended product combination: After the algorithm runs, the optimal recommended product combination that meets the conditions is output.

[0119] Step S404, verifying the optimal recommended product combination, and after the verification is passed, outputting the preset profit prediction model and the simulated annealing algorithm combination as the profit-risk balance model.

[0120] In this embodiment, the verification of the optimal recommended product portfolio is achieved through historical data backtesting, market simulation and other methods, wherein historical data backtesting is a method for evaluating the performance of a specific investment strategy or product portfolio in the past market environment by simulating historical transactions. The verification of the optimal recommended product portfolio through historical data backtesting includes the following steps: Data collection: First, it is necessary to collect historical data covering multiple market cycles (such as bull market, bear market, shock market, etc.). These data should include stock prices, trading volume, macroeconomic indicators, etc. Strategy construction: According to investment goals and risk tolerance, construct a product portfolio or investment strategy to be verified. For example, a balanced investment portfolio containing different asset categories such as stocks, bonds, and commodities can be designed. Backtest execution: Use historical data to simulate transactions according to the constructed investment strategy. Record the buying price, selling price, transaction cost and other information of each transaction, and calculate the total return, annualized rate of return, maximum drawdown and other key indicators of the investment portfolio. Result analysis: According to the backtest results, analyze the performance of the investment strategy in different market environments. Evaluate its profitability, risk control ability, market adaptability, etc. Market simulation is a method of simulating actual market trading behavior by constructing a virtual market environment. Its purpose is to test the performance of investment strategies or product portfolios under specific market conditions and evaluate their adaptability and robustness. The verification of the optimal recommended product portfolio through market simulation includes the following steps: Environment construction: Build a virtual market environment based on historical data and current market conditions. Including setting market rules, trading mechanisms, participant behaviors, etc. Strategy deployment: Deploy the investment strategy or product portfolio to be verified into the virtual market environment. Simulated trading: Conduct simulated trading in the virtual market environment, record trading data and calculate relevant indicators. Result evaluation: Evaluate the performance of investment strategies or product portfolios based on the results of simulated trading. Including profitability, risk control ability, market adaptability, etc. Finally, the combination of the preset return prediction model and the simulated annealing algorithm is output as a return-risk balance model. This model can be used as an independent tool or module for product portfolio recommendation and risk-return evaluation.

[0121] This embodiment obtains the product combination rate of return by inputting the recommended product combination into the preset profit prediction model for profit calculation; obtains the preset expected rate of return, and screens the recommended product combination according to the preset expected rate of return and the product combination rate of return to obtain the expected recommended product combination; calculates the optimal solution for the expected recommended product combination based on the simulated annealing algorithm to obtain the optimal recommended product combination; verifies the optimal recommended product combination, and after verification, outputs the preset profit prediction model and the simulated annealing algorithm combination as the profit-risk balance model. Thus, a profit-risk balance model that can accurately output the optimal recommended product under profit and risk considerations is obtained, so as to facilitate the subsequent sorting operation of the user's preferred product combination.

[0122] Continue to refer Figure 7 In some optional implementations of this embodiment, step S50 includes the following steps:

[0123] Step S501, obtaining user historical purchase data according to the user portrait;

[0124] In this embodiment, the corresponding user financial data is obtained based on the user portrait, and a search is performed in the user financial data to obtain the user's historical purchase data. The user's historical purchase data includes the type of purchased products, price, purchase time, purchase frequency, return rate, evaluation information, etc. The user's historical purchase data can be obtained from e-commerce platforms, CRM systems, payment records and other channels.

[0125] Step S502, parsing the user's historical purchase data to obtain the user's historical purchased products and the user's purchase amount;

[0126] In this embodiment, by performing a detailed analysis on the acquired historical purchase data, a list of specific products purchased by the user (i.e., the user's historically purchased products) and the purchase amount of each product are extracted, thereby obtaining the user's historically purchased products and the user's product purchase amount.

[0127] Step S503, constructing the purchase preference model according to the user's historical purchases of products and the amount of money the user spent on the products;

[0128] In this embodiment, the parsed user's historical purchase products and corresponding purchase amounts are used as input to construct a user's purchase preference model through statistical analysis, machine learning algorithms (such as association rule mining, cluster analysis, collaborative filtering, etc.) or other technologies. This model can reflect the user's preference for different types of products. In this embodiment, the construction of a purchase preference model through statistical analysis includes the following steps: Frequency analysis: Count the number of times users purchase each type of product to understand the user's purchase frequency and preference direction. Cross-tab analysis: Analyze the purchase correlation between different products and identify product combinations that users often purchase together. Trend analysis: Observe the changing trend of user purchase behavior over time and predict the user's future purchase needs. Through descriptive statistics, correlation analysis and other means of user historical purchase data, the general laws and characteristics of user purchase behavior are revealed, and the construction of a purchase preference model is realized in combination with a machine learning algorithm model.

[0129] Step S504, inputting the user's preferred product combination into the purchase preference model to obtain a first model score;

[0130] In this embodiment, the first model score is a score that reflects the degree of match between the user's preferred product combination and the user's purchasing preference.

[0131] Step S505, inputting the user's preferred product portfolio into the return-risk balance model to obtain a second model score;

[0132] In this embodiment, the second model score is a score that reflects the balance between the return and risk of the user's preferred product portfolio.

[0133] Step S506, calculating a comprehensive model score according to the first model score and the second model score;

[0134] In this embodiment, according to the first model score (purchase preference matching) and the second model score (return-risk balance), the model comprehensive score is obtained by weighted calculation, and the calculation formula of the model comprehensive score is Z=a*X+b*Y, where X is the total model comprehensive score, a is the weight corresponding to the first model score, X is the first model score, b is the weight corresponding to the second model score, and Y is the second model score. In this embodiment, a and b are preset to 0.5, and can be adjusted accordingly according to actual conditions.

[0135] Step S507, sorting the user preferred product combinations according to the comprehensive score of the model to obtain the product combination sorting table.

[0136] In this embodiment, the product combination ranking table is sorted from high to low according to the comprehensive model scores corresponding to the user's preferred product combinations, and can also be sorted according to the order set according to actual needs, such as the amount of the user's preferred product combinations.

[0137] This embodiment obtains the user's historical purchase data according to the user portrait; parses the user's historical purchase data to obtain the user's historically purchased products and the user's product purchase amount; constructs the purchase preference model according to the user's historically purchased products and the user's product purchase amount; inputs the user's preferred product combination into the purchase preference model to obtain a first model score; inputs the user's preferred product combination into the benefit-risk balance model to obtain a second model score; calculates a comprehensive model score according to the first model score and the second model score; and sorts the user's preferred product combination according to the model comprehensive score, thereby obtaining a product combination ranking table based on the user's purchase preference and benefit-risk balance considerations, so as to facilitate subsequent product combination recommendation operations for users.

[0138] In some optional implementations of this embodiment, after step S60, the following steps are further included:

[0139] Collect product push feedback data;

[0140] In this embodiment, after the product is pushed to the user, the system needs to collect the user's feedback data on the pushed product combination. Such feedback data may include whether the user clicks on the push, whether the user views the product combination details, whether the user adds the product combination to the purchase list, whether the user finally purchases the product combination, and the user's evaluation of the product combination.

[0141] Calculating a deviation value between the product push feedback data and the expected feedback data of the product combination ranking table;

[0142] In this embodiment, the expected feedback data can be obtained based on historical data, model predictions, and other prior knowledge, or can be preset by data processing personnel according to actual conditions. The deviation value is a comprehensive indicator reflecting the expected deviation, such as click-through rate deviation, purchase rate deviation, satisfaction deviation, etc., which can be analyzed and calculated by comparing the product push feedback data and the expected feedback data after the product combination ranking table is pushed.

[0143] Determine whether the deviation value is greater than or equal to a preset threshold;

[0144] In this embodiment, the preset threshold is used to determine whether the deviation between the actual feedback and the expected feedback is within an acceptable range. The preset threshold can be determined based on factors such as business requirements, user expectations, and system performance.

[0145] If the deviation value is greater than or equal to the preset threshold, the purchase preference model and the return-risk balance model are evaluated according to the deviation value, and the purchase preference model and the return-risk balance model are optimized according to the evaluation result to obtain the optimized purchase preference model and the optimized return-risk balance model, and a new product combination ranking table is regenerated to push product combinations to the user until the deviation value corresponding to the new product combination ranking table is less than the preset threshold;

[0146] In this embodiment, the purchase preference model and the benefit-risk balance model are optimized according to the evaluation results to obtain the optimized purchase preference model and the optimized benefit-risk balance model, and a new product combination ranking table is regenerated to push product combinations to users as an iterative optimization step. By executing the iterative optimization step, it is effectively ensured that the push of the product combination ranking table obtains the expected results, so as to effectively improve the accuracy of product combination push and user satisfaction.

[0147] If the deviation value is less than the preset threshold, the product push feedback data is recorded in the result feedback record table for storage.

[0148] In this embodiment, the result feedback record table is saved in a local folder of the financial business system, and the product push feedback data can be sent to the result feedback record table for storage through a corresponding path.

[0149] This embodiment collects product push feedback data; calculates the deviation value between the product push feedback data and the expected feedback data of the product combination ranking table; determines whether the deviation value is greater than or equal to a preset threshold; if the deviation value is greater than or equal to the preset threshold, the purchase preference model and the benefit-risk balance model are evaluated according to the deviation value, and the purchase preference model and the benefit-risk balance model are optimized according to the evaluation result to obtain an optimized purchase preference model and an optimized benefit-risk balance model, and a new product combination ranking table is regenerated to push product combinations to users until the deviation value corresponding to the new product combination ranking table is less than the preset threshold; if the deviation value is less than the preset threshold, the product push feedback data is recorded in the result feedback record table for storage. This effectively ensures that the product recommendations made to users are accurate and effective, so as to improve the effect of product combination recommendations.

[0150] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0151] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0152] Further references Figure 8 , as a response to the above Figure 1 The present application provides an embodiment of a financial product recommendation device, which is similar to Figure 1 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0153] like Figure 8As shown, the financial product recommendation device 700 described in this embodiment includes: a data acquisition module 701, a product combination acquisition module 702, a product combination screening module 703, a model training module 704, a product combination sorting module 705, and a product combination pushing module 706.

[0154] in:

[0155] Data acquisition module 701, used to acquire real-time financial data and user portraits;

[0156] A product portfolio acquisition module 702 is used to build a risk preference model based on the user portrait and acquire the user's preferred product portfolio according to the risk preference model;

[0157] The product combination screening module 703 is used to obtain the expected rate of return, and screen the user's preferred product combination according to the expected rate of return to obtain a recommended product combination;

[0158] A model training module 704 is used to train a preset revenue prediction model according to the recommended product combination to obtain a revenue-risk balance model;

[0159] A product combination ranking module 705 is used to build a purchase preference model based on the user portrait, and to rank the user's preferred product combinations according to the purchase preference model and the benefit-risk balance model to obtain a product combination ranking table;

[0160] The product combination push module 706 is used to push product combinations to users according to the product combination ranking table.

[0161] By adopting the above-mentioned financial product recommendation device, this embodiment can obtain real-time financial data and user portraits; build a risk preference model based on the user portrait, and obtain the user's preferred product portfolio according to the risk preference model; obtain the expected rate of return, screen the user's preferred product portfolio according to the expected rate of return, and obtain the recommended product portfolio; train the preset income prediction model according to the recommended product portfolio to obtain the income-risk balance model; build a purchase preference model based on the user portrait, sort the user's preferred product portfolio according to the purchase preference model and the income-risk balance model, and obtain a product portfolio ranking table; push product portfolios to users according to the product portfolio ranking table. This effectively realizes personalized and accurate financial product portfolio recommendations for users, so as to provide users with accurate and effective investment advice and references.

[0162] To solve the above technical problems, the present application also provides a computer device. Fig. 9 , Fig. 9 This is a basic structural block diagram of the computer device in this embodiment.

[0163] The computer device 8 includes a memory 81, a processor 82, and a network interface 83 that are interconnected through a system bus. It should be noted that the figure only shows a computer device 8 with components 81-83, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (Application Specific Integrated Circuit, ASIC), programmable gate arrays (Field-Programmable Gate Array, FPGA), digital processors (Digital Signal Processor, DSP), embedded devices, etc.

[0164] The computer device may be a computing device such as a desktop computer, a notebook, a PDA, a cloud server, etc. The computer device may interact with a user through a keyboard, a mouse, a remote controller, a touch pad, or a voice control device.

[0165] The memory 81 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 81 can be an internal storage unit of the computer device 8, such as a hard disk or memory of the computer device 8. In other embodiments, the memory 81 can also be an external storage device of the computer device 8, such as a plug-in hard disk, a smart memory card (Smart Media Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (FlashCard), etc. equipped on the computer device 8. Of course, the memory 81 can also include both the internal storage unit of the computer device 8 and its external storage device. In this embodiment, the memory 81 is generally used to store the operating system and various application software installed on the computer device 8, such as computer-readable instructions of the financial product recommendation method, etc. In addition, the memory 81 can also be used to temporarily store various types of data that have been output or are to be output.

[0166] The processor 82 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor 82 is generally used to control the overall operation of the computer device 8. In this embodiment, the processor 82 is used to run computer-readable instructions or process data stored in the memory 81, such as computer-readable instructions for running the financial product recommendation method.

[0167] The network interface 83 may include a wireless network interface or a wired network interface. The network interface 83 is generally used to establish a communication connection between the computer device 8 and other electronic devices.

[0168] By adopting the above-mentioned computer equipment, this embodiment can obtain real-time financial data and user portraits; build a risk preference model based on the user portrait, and obtain the user's preferred product portfolio according to the risk preference model; obtain the expected rate of return, screen the user's preferred product portfolio according to the expected rate of return, and obtain the recommended product portfolio; train the preset return prediction model according to the recommended product portfolio to obtain a return-risk balance model; build a purchase preference model based on the user portrait, sort the user's preferred product portfolio according to the purchase preference model and the return-risk balance model, and obtain a product portfolio ranking table; push product portfolios to users according to the product portfolio ranking table. This effectively realizes personalized and accurate financial product portfolio recommendations for users, so as to provide users with accurate and effective investment advice and references.

[0169] The present application also provides another implementation, namely, providing a computer-readable storage medium, wherein the computer-readable storage medium stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the financial product recommendation method as described above.

[0170] By adopting the above-mentioned computer-readable storage medium, this embodiment can obtain real-time financial data and user portraits; build a risk preference model based on the user portrait, and obtain the user's preferred product portfolio according to the risk preference model; obtain the expected rate of return, screen the user's preferred product portfolio according to the expected rate of return, and obtain the recommended product portfolio; train the preset return prediction model according to the recommended product portfolio to obtain a return-risk balance model; build a purchase preference model based on the user portrait, sort the user's preferred product portfolio according to the purchase preference model and the return-risk balance model, and obtain a product portfolio ranking table; push product portfolios to users according to the product portfolio ranking table. This effectively realizes personalized and accurate financial product portfolio recommendations for users, so as to provide users with accurate and effective investment advice and references.

[0171] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in each embodiment of the present application.

[0172] Obviously, the embodiments described above are only some embodiments of the present application, rather than all embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific implementation methods, or to perform equivalent replacement of some of the technical features therein. Any equivalent structure made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, is similarly within the scope of patent protection of this application.

Claims

1. A method for recommending financial products, characterized in that: The steps include: Obtain real-time financial data and user portraits; Building a risk preference model based on the user portrait, and obtaining the user's preferred product portfolio according to the risk preference model; Obtaining an expected rate of return, and screening the user's preferred product portfolio according to the expected rate of return to obtain a recommended product portfolio; Training a preset revenue prediction model according to the recommended product portfolio to obtain a revenue-risk balance model; Building a purchase preference model based on the user portrait, and sorting the user's preferred product portfolio according to the purchase preference model and the benefit-risk balance model to obtain a product portfolio sorting table; Product combinations are pushed to users according to the product combination ranking table.

2. The financial product recommendation method according to claim 1, characterized in that: The steps of obtaining real-time financial data and user portraits specifically include: Extracting the financial real-time data from a financial database in real time; Obtaining user collected information, and organizing and analyzing the user collected information to obtain user key characteristics and user trends; The user portrait is constructed according to the user key features and the user trends.

3. The financial product recommendation method according to claim 1, characterized in that: The step of building a risk preference model based on the user portrait and obtaining the user's preferred product portfolio according to the risk preference model specifically includes: Acquire the user's historical purchase data and psychological evaluation data based on the user portrait; Constructing the risk preference model based on the user's historical purchase data and the psychological assessment data; Products that meet the user's risk level are screened and combined according to the risk preference model to obtain the user's preferred product combination.

4. The financial product recommendation method according to claim 1, characterized in that: The step of obtaining the expected rate of return, screening the user's preferred product combination according to the expected rate of return, and obtaining the recommended product combination specifically includes: Obtain user account information and market product data; Acquire the expected rate of return according to the user account information; The user preferred product combination is screened according to the expected rate of return and the market product data to obtain the recommended product combination.

5. The financial product recommendation method according to claim 1, characterized in that: The step of training a preset revenue prediction model according to the recommended product combination to obtain a revenue-risk balance model specifically includes: Inputting the recommended product combination into the preset profit prediction model to perform profit calculation to obtain the product combination rate of return; Obtaining a preset expected rate of return, and screening the recommended product combination according to the preset expected rate of return and the product combination rate of return to obtain an expected recommended product combination; Calculating the optimal solution for the expected recommended product combination based on a simulated annealing algorithm to obtain an optimal recommended product combination; The optimal recommended product combination is verified, and after the verification is passed, the preset profit prediction model and the simulated annealing algorithm are combined and output as the profit-risk balance model.

6. The financial product recommendation method according to claim 1, characterized in that: The step of constructing a purchase preference model based on the user portrait, and sorting the user preferred product combinations according to the purchase preference model and the benefit-risk balance model to obtain a product combination sorting table specifically includes: Acquire the user's historical purchase data according to the user portrait; Parsing the user's historical purchase data to obtain the user's historically purchased products and the user's purchase amount; Constructing the purchase preference model based on the user's historical purchases of products and the amount of money the user spent on the products; Inputting the user's preferred product combination into the purchase preference model to obtain a first model score; Inputting the user's preferred product portfolio into the return-risk balance model to obtain a second model score; Calculate a comprehensive model score according to the first model score and the second model score; The user preferred product combinations are sorted according to the comprehensive scores of the model to obtain a product combination sorting table.

7. The financial product recommendation method according to claim 1, characterized in that: After the step of pushing product combinations to users according to the product combination ranking table, the following steps are also included: Collect product push feedback data; Calculating a deviation value between the product push feedback data and the expected feedback data of the product combination ranking table; Determine whether the deviation value is greater than or equal to a preset threshold; If the deviation value is greater than or equal to the preset threshold, the purchase preference model and the return-risk balance model are evaluated according to the deviation value, and the purchase preference model and the return-risk balance model are optimized according to the evaluation results to obtain the optimized purchase preference model and the optimized return-risk balance model, and a new product combination ranking table is regenerated to push product combinations to users until the deviation value corresponding to the new product combination ranking table is less than the preset threshold.

8. A financial product recommendation device, characterized in that: include: Data acquisition module, used to obtain real-time financial data and user portraits; A product portfolio acquisition module, used to build a risk preference model based on the user portrait, and acquire the user's preferred product portfolio according to the risk preference model; A product combination screening module is used to obtain an expected rate of return, and screen the user's preferred product combination according to the expected rate of return to obtain a recommended product combination; A model training module is used to train a preset revenue prediction model according to the recommended product portfolio to obtain a revenue-risk balance model; A product combination ranking module, used to build a purchase preference model based on the user portrait, and to sort the user preferred product combinations according to the purchase preference model and the benefit-risk balance model to obtain a product combination ranking table; The product combination push module is used to push product combinations to users according to the product combination ranking table.

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

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

Images