Intelligent financial product matching method and system

Abstract

The invention relates to the technical field of financial science and technology, and discloses an intelligent financial product matching method and system, and the method comprises the steps: analyzing occupational attributes and income levels in user identity data, and constructing an exclusive financial demand portrait of a financial user; screening a candidate financial product set meeting a fund flow constraint condition; calculating a matching preference coupling degree between the risk preference tag and the risk level identifier, calculating a cooperation index between the financial planning reference and the income cycle parameter, and generating a multi-dimensional matching index of the candidate financial product set; calculating a settlement prediction value of the candidate financial product set, and generating a financial product recommendation sequence of the candidate financial product set; and performing optimization processing on the financial product recommendation sequence to obtain an optimized product recommendation sequence, and generating a product recommendation set of the financial users. The accuracy of financial product matching can be improved.

Description

Technical Field

[0001] The present invention relates to a method and system for intelligent financial product matching, and belongs to the technical field of financial technology. Background Art

[0002] As global financial markets continue to expand and financial products become increasingly diversified, investors' demands for financial products are becoming increasingly complex and diverse. From individual investors to corporate institutions, everyone hopes to achieve wealth appreciation, risk diversification, and meet specific financial goals through the rational allocation of financial products. In order to accurately match investor needs with a large number of financial products, intelligent financial product matching methods and systems have emerged.

[0003] Traditional financial product matching methods mainly rely on manual experience and simple questionnaire assessments. Financial advisors manually screen financial products and recommend them to clients based on their own experience and limited understanding of the clients. Under this method, financial advisors need to spend a lot of time and energy to familiarize themselves with the characteristics of various financial products. At the same time, due to the limitations of personal cognition and energy, it is difficult to fully and accurately match client needs. Questionnaire assessments are often designed to be relatively simple and can only obtain some basic financial information and risk preferences of clients. They cannot gain in-depth insights into the dynamic needs of clients in different economic environments, nor the comprehensive performance of financial products under complex market conditions.

[0004] With the explosive growth of financial market data, including macroeconomic data, market data, and historical performance data of financial products, traditional matching methods have difficulty in real-time and efficient analysis and utilization of these massive amounts of data. In addition, the financial market environment is changing rapidly, and factors such as interest rate fluctuations, policy adjustments, and industry changes frequently affect the returns and risks of financial products. Traditional methods are unable to adjust matching strategies in a timely manner according to market dynamics, which leads to poor adaptability of traditional financial product matching methods, and in turn reduces the accuracy of financial product matching. Summary of the Invention

[0005] The present invention provides an intelligent financial product matching method and system, the main purpose of which is to improve the accuracy of financial product matching.

[0006] To achieve the above objectives, the present invention provides an intelligent financial product matching method, comprising: Obtaining the user identity data and historical settlement records of the financial user, parsing the occupational attributes and income level in the user identity data, extracting the risk preference labels and consumption behavior characteristics in the historical settlement records, and combining the occupational attributes, income level, and risk preference labels to construct a unique financial needs profile of the financial user; Collect the financial planning benchmarks and funds flow constraints input by the financial user in real time, combine them with a preset financial market product database and the financial demand profile, screen a set of candidate financial products that meet the funds flow constraints, and identify the risk level identifiers and return cycle parameters in the candidate financial product set; Calculating the matching preference coupling degree between the risk preference label and the risk level identifier, calculating the synergy index between the financial planning benchmark and the revenue cycle parameter, and combining the matching preference coupling degree and the synergy index to generate a multi-dimensional matching index for the candidate financial product set; Monitoring rate fluctuation data and credit risk indicators in the financial market of the financial user, calculating settlement prediction values ​​for the candidate financial product set based on the rate fluctuation data and the credit risk indicators, and generating a financial product recommendation sequence for the candidate financial product set based on the settlement prediction values ​​and the multi-dimensional matching index; Extract the frequency of money use and historical redemption operation records from the consumption behavior characteristics, combine the frequency of money use and the historical redemption operation records, optimize the financial product recommendation sequence, and obtain an optimized product recommendation sequence. Based on the optimized product recommendation sequence and the set of candidate financial products, generate a product recommendation set for the financial user.

[0007] Optionally, extracting the risk preference labels and consumption behavior characteristics from the historical settlement records includes: Cleaning the historical settlement records to obtain target settlement records; extracting record features from the target settlement record, and performing clustering processing on the record features to obtain clustered record features; Analyzing the characteristic behavior tags corresponding to the cluster record features, performing risk scoring processing on the characteristic behavior tags to obtain a risk score value; Determining a risk preference label in the characteristic behavior label based on the risk score value; Based on the characteristic behavior labels, consumption behavior features of the cluster record features are extracted.

[0008] Optionally, the combining of the occupational attributes, the income level, and the risk preference label to construct a financial demand profile specific to the financial user includes: Performing standard classification processing on the occupational attributes to obtain standard occupational categories; Performing quantitative grading processing on the income levels to obtain quantitative income grades; Performing weight assignment on the risk preference labels to obtain label weights; fusing the standard occupational category, the income quantification level, and the label weight to obtain a financial demand feature vector of the financial user; determining the financial demand characteristics of the financial user based on the financial demand characteristic vector; The financial demand characteristics are personalized and adapted to obtain a financial demand profile that is exclusive to the financial user.

[0009] Optionally, the combining of a preset financial market product database and the financial demand profile to screen a set of candidate financial products that meet the funds flow constraint condition includes: Extracting an identifier from the financial demand profile to obtain a financial demand identifier; Extracting product feature identifiers from the financial market product database; calculating an identifier matching degree between the financial demand identifier and the product feature identifier, and extracting an initial financial product from the financial market product database based on the identifier matching degree; The initial financial products are subjected to yield adaptation processing to obtain a set of candidate financial products meeting the fund flow constraint condition.

[0010] Optionally, the calculating the identifier matching degree between the financial demand identifier and the product feature identifier includes: Performing vectorization processing on the financial demand identifier and the product feature identifier to obtain a demand identifier vector and a feature identifier vector; Calculate the vector matching degree between the requirement identification vector and the feature identification vector: ; in, Indicates the vector matching degree between the requirement identification vector and the feature identification vector, represents the ath vector in the demand identification vector, represents the bth vector in the feature identification vector, a and b represent the serial numbers corresponding to the requirement identification vector and the feature identification vector respectively, and m and n represent the number corresponding to the requirement identification vector and the feature identification vector respectively; Based on the vector matching degree, an identification matching degree between the financial demand identification and the product feature identification is obtained.

[0011] Optionally, the calculating the matching preference coupling degree between the risk preference label and the risk level identifier includes: Encoding the risk preference label and the risk level identifier to obtain a risk preference coding value and a risk level coding value; Normalizing the risk preference code value and the risk level code value to obtain a normalized preference code value and a normalized level code value; In combination with the normalized preference code value and the normalized grade code value, the matching preference coupling degree between the risk preference label and the risk grade identifier is calculated by the following formula: ; in, Indicates the matching preference coupling between the risk preference label and the risk level identifier, represents the normalized preference encoding value, represents the normalized rank-coded value, Indicates selecting the maximum value between the normalized preference coding value and the normalized rank coding value.

[0012] Optionally, the calculating a synergy index between the financial planning benchmark and the revenue cycle parameter includes: Performing planning target analysis on the financial planning benchmark to obtain a financial target time domain; Performing segmented quantization processing on the revenue cycle parameter to obtain segmented cycle values; The matching degree between the financial target time domain and the period segment value is calculated to obtain the time matching degree: The matching degree between the financial target time domain and the period segment value is calculated using the following formula: ; Performing weight distribution processing on the time matching degree to obtain a matching degree weight; The synergy index between the financial planning benchmark and the revenue cycle parameter is calculated by combining the matching weight and the time matching.

[0013] Optionally, the calculating the settlement prediction value of the candidate financial product set by combining the rate fluctuation data and the credit risk indicator includes: Extracting a time series rate value from the rate fluctuation data, and smoothing the time series rate value to obtain a smoothed time series rate value; Calculating a rate trend coefficient corresponding to the rate fluctuation data based on the smoothed time-series rate value; Calculating the risk-adjusted value corresponding to the credit risk indicator; The settlement forecast value of the candidate financial product set is calculated by combining the rate trend coefficient and the risk adjustment value using the following formula: ; in, represents the settlement forecast value of the candidate financial product set, Indicates the benchmark settlement value, represents the rate trend coefficient, Represents the risk-adjusted value.

[0014] Optionally, optimizing the financial product recommendation sequence based on the frequency of fund usage and the historical redemption operation records to obtain an optimized product recommendation sequence includes: Extracting the redemption operation behavior and the redemption operation object from the historical redemption operation record; Analyzing the coupling mechanism between the redemption operation behavior and the redemption operation object; Sorting the usage frequencies of the funds to obtain sorted usage frequencies of the funds; The financial product recommendation sequence is optimized by combining the coupling mechanism and the usage frequency of the ranked items to obtain an optimized product recommendation sequence.

[0015] In order to solve the above problems, the present invention also provides an intelligent financial product matching system, which includes: A financial needs profile building module is used to obtain the user identity data and historical settlement records of a financial user, analyze the occupational attributes and income level in the user identity data, extract the risk preference labels and consumption behavior characteristics in the historical settlement records, and combine the occupational attributes, income level and risk preference labels to build a financial needs profile unique to the financial user; A financial product screening module is configured to collect the financial planning benchmarks and cash flow constraints input by the financial user in real time, and screen a set of candidate financial products that meet the cash flow constraints based on a preset financial market product database and the financial demand profile, and identify the risk level identifiers and return cycle parameters in the set of candidate financial products; a multidimensional matching index generation module, configured to calculate a matching preference coupling degree between the risk preference label and the risk level identifier, calculate a synergy index between the financial planning benchmark and the revenue cycle parameter, and generate a multidimensional matching index for the candidate financial product set by combining the matching preference coupling degree and the synergy index; a financial product recommendation sequence generation module, configured to monitor rate fluctuation data and credit risk indicators in the financial market of the financial user, calculate settlement prediction values ​​for the candidate financial product set based on the rate fluctuation data and the credit risk indicators, and generate a financial product recommendation sequence for the candidate financial product set based on the settlement prediction values ​​and the multi-dimensional matching index; The recommendation sequence optimization module is used to extract the frequency of payment usage and historical redemption operation records from the consumption behavior characteristics, and optimize the financial product recommendation sequence based on the frequency of payment usage and the historical redemption operation records to obtain an optimized product recommendation sequence. Based on the optimized product recommendation sequence and the set of candidate financial products, a product recommendation set for the financial user is generated.

[0016] Compared with the problems described in the background technology, the present invention can gain insight into the user's economic background and professional characteristics by analyzing the occupational attributes and income levels in the user's identity data, and provide a basis for the positioning of financial services; extracting the risk preference labels and consumption behavior characteristics in the historical settlement records can accurately grasp the user's investment tendencies and consumption habits, and provide an important basis for the subsequent construction of the financial demand portrait exclusive to the financial user. Furthermore, the present invention can locate financial products that meet the user's financial status, capital flow needs and risk preferences by combining a preset financial market product database with the financial demand portrait, thereby greatly reducing the time cost of users screening among massive products and improving the accuracy and adaptability of financial services. The present invention calculates the risk preference labels and the risk The matching preference coupling degree between the grade identifiers can clearly show the degree of fit between the user's subjective risk-bearing willingness and the objective risk level of the financial product, helping users to quickly identify financial products that match their own risk preferences, reduce investment risks, and improve the rationality of investment decisions. Furthermore, the present invention calculates the settlement forecast value of the candidate financial product set by combining the rate fluctuation data and the credit risk indicator. The estimated return of each product in the candidate financial product set can be understood through the settlement forecast value, thereby improving the generation accuracy of the financial product recommendation sequence of the candidate financial product set. Furthermore, the present invention optimizes the financial product recommendation sequence by combining the frequency of use of the funds and the historical redemption operation records, generates a personalized product recommendation set, and improves the practicality of financial product recommendations and user satisfaction. Therefore, the intelligent financial product matching method and system provided in the embodiment of the present invention can improve the accuracy of financial product matching. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] Figure 1 A flowchart of an intelligent financial product matching method provided by one embodiment of the present invention; Figure 2 A schematic diagram of modules for implementing the intelligent financial product matching method provided in one embodiment of the present invention.

[0018] The purpose, features and advantages of the present invention will be further described with reference to the accompanying drawings and in conjunction with the embodiments. DETAILED DESCRIPTION

[0019] It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention.

[0020] The present application provides a method for matching intelligent financial products. The method can be executed by at least one of a server, a terminal, or other electronic device capable of executing the method provided in the present application. In other words, the method can be executed by software or hardware installed on a terminal or server device. The server can include, but is not limited to, a single server, a server cluster, a cloud server, or a cloud server cluster.

[0021] Example 1: Reference Figure 1 FIG. 1 is a flow chart of a smart financial product matching method according to an embodiment of the present invention. In this embodiment, the smart financial product matching method includes: S1. Obtain the user identity data and historical settlement records of the financial user, parse the occupational attributes and income level in the user identity data, extract the risk preference labels and consumption behavior characteristics in the historical settlement records, and combine the occupational attributes, income level and risk preference labels to construct a financial demand profile exclusive to the financial user.

[0022] By parsing the occupational attributes and income levels in the user identity data, the present invention can provide insights into the user's economic background and professional characteristics, providing a basis for positioning financial services. By extracting the risk preference labels and consumption behavior characteristics from the historical settlement records, the user's investment tendencies and consumption habits can be accurately grasped, providing an important basis for the subsequent construction of a financial demand profile unique to the financial user. It should be explained that the user identity data is the personal information provided by the financial user when registering or handling business, typically including occupation, income, age, gender, etc.; the historical settlement records are records of the financial user's capital flow in past transactions, such as consumption records, investment records, and loan records; the occupational attributes are the user's occupation type, such as corporate employee, freelancer, civil servant, etc.; the income level is the user's income level classification, such as high income, middle income, low income, etc.; the risk preference label is the user's risk tolerance for financial products, such as conservative, stable, aggressive, etc.; the consumption behavior characteristics are the behavioral patterns exhibited by the user during the consumption process, such as consumption frequency, consumption amount, consumption category, etc.; further, the parsing of the occupational attributes and income level in the user identity data can be achieved through natural language processing technology.

[0023] Specifically, the extraction of risk preference labels and consumption behavior characteristics from the historical settlement records includes: Cleaning the historical settlement records to obtain target settlement records; extracting record features from the target settlement record, and performing clustering processing on the record features to obtain clustered record features; Analyzing the characteristic behavior tags corresponding to the cluster record features, performing risk scoring processing on the characteristic behavior tags to obtain a risk score value; Determining a risk preference label in the characteristic behavior label based on the risk score value; Based on the characteristic behavior labels, consumption behavior features of the cluster record features are extracted.

[0024] It should be explained that the target settlement record is a specific part of the historical settlement record that has been screened and used for key analysis; the record feature is the key data attribute in the target settlement record, such as transaction amount, transaction time, etc.; the cluster record feature is a set of features with similar characteristics formed after cluster analysis of the record feature; the characteristic behavior label is an identifier corresponding to the cluster record feature that can reflect the user's specific behavior pattern; the risk score value is a quantitative value calculated by the characteristic behavior label according to the pre-set risk assessment rules; the risk preference label is a label that is screened and determined from the characteristic behavior label based on the risk score value to represent the user's risk tendency; the consumption behavior feature is the key information that can reflect the user's daily consumption habits extracted from the cluster record feature based on the characteristic behavior label, such as consumption frequency, proportion of consumption areas, etc.

[0025] Furthermore, the historical settlement records can be cleaned by data cleaning algorithms such as data denoising and missing value filling to obtain target settlement records; the record features in the target settlement records can be extracted by the attribute extraction method in data mining technology, and the record features can be clustered by clustering algorithms such as K-Means to obtain cluster record features; the characteristic behavior labels corresponding to the cluster record features can be analyzed by combining expert experience with a behavioral pattern analysis model, and the characteristic behavior labels can be risk-scored by a preset risk assessment index system to obtain a risk score value. The risk assessment index system can be constructed by comprehensively considering financial industry standards, risk factor analysis in historical data, expert experience, and in-depth research on various financial products and market environments; based on the risk score value, the risk preference label in the characteristic behavior label is determined according to the set risk preference division threshold; the consumption behavior characteristics of the cluster record characteristics can be extracted from the characteristic behavior labels by screening rules for consumption-related characteristics.

[0026] The present invention constructs a financial demand profile exclusive to the financial user by combining the occupational attributes, the income level and the risk preference label, which can understand the potential financial needs of the financial user and help financial institutions customize personalized products and services. It should be explained that the financial demand profile is a comprehensive portrait of the personalized financial needs exclusive to the financial user.

[0027] Specifically, the combination of the occupational attributes, the income level, and the risk preference label to construct a financial demand profile exclusive to the financial user includes: Performing standard classification processing on the occupational attributes to obtain standard occupational categories; Performing quantitative grading processing on the income levels to obtain quantitative income grades; Performing weight assignment on the risk preference labels to obtain label weights; fusing the standard occupational category, the income quantification level, and the label weight to obtain a financial demand feature vector of the financial user; determining the financial demand characteristics of the financial user based on the financial demand characteristic vector; The financial demand characteristics are personalized and adapted to obtain a financial demand profile that is exclusive to the financial user.

[0028] It should be explained that the standard occupational category is a clear occupational category obtained after the occupational attributes are subjected to standard classification processing and divided according to the general occupational classification system; the income quantification level is a level identification obtained after the income level is quantitatively graded and quantified according to the set income range and relevant economic indicators; the label weight is the numerical weight assigned to the risk preference label based on the degree of its influence on the user's financial decision-making; the financial demand feature vector is obtained by fusing the standard occupational category, the income quantification level and the label weight, and is a quantitative vector of the financial user that can comprehensively reflect the user's financial characteristics, covering multi-dimensional information such as occupational economic characteristics, income strength and risk tendency; the financial demand characteristics are the key elements of the financial user represented by the financial demand feature vector, which reflect their unique characteristics in terms of financial product selection, investment strategy tendencies and consumer financial needs.

[0029] Furthermore, the occupational attributes can be subjected to standard classification processing by referring to the occupational classification standard library to obtain standard occupational categories; the income levels can be quantitatively graded by setting clear income intervals and combining statistical analysis methods to obtain income quantification levels. For example, based on local resident income statistics and industry average levels, annual income can be divided into five intervals: low, medium-low, medium, medium-high, and high, corresponding to levels 1 to 5 respectively, to intuitively show the user's income level position in the overall economic environment; the risk preference and financial decision-making correlation analysis in historical data can be combined to assign weights to the risk preference labels according to the correlation of the analysis to obtain label weights; the standard occupational categories can be classified by vector space models. The financial demand feature vector of the financial user is obtained by fusing the income quantification level and the label weight, such as the Word2Vecxx model; based on the financial demand feature vector, the financial demand characteristics of the financial user are determined by using a feature extraction and interpretation algorithm, such as analyzing the values ​​of each dimension of the vector through a decision tree algorithm to identify key features that have a significant impact on the user's investment product selection, consumer credit tendency, etc., so as to determine the financial demand characteristics; the financial demand characteristics are personalized and adapted to obtain a financial demand portrait exclusive to the financial user, such as presenting the user's personalized financial needs in terms of investment product type preference, credit limit demand, financial planning tendency, etc. in the form of a visual chart.

[0030] S2. Collect the financial planning benchmarks and funds flow constraints input by the financial user in real time, combine them with a preset financial market product database and the financial demand profile, screen a set of candidate financial products that meet the funds flow constraints, and identify the risk level identifier and return cycle parameters in the candidate financial product set.

[0031] The present invention combines a preset financial market product database with the financial demand portrait to screen a set of candidate financial products that meet the cash flow constraints, thereby locating financial products that meet the user's financial status, cash flow needs and risk preferences, greatly reducing the time cost of users screening among massive products, and improving the accuracy and adaptability of financial services. It should be explained that the financial planning benchmark is a key indicator set by the user based on his or her own financial goals, such as the expected asset appreciation rate, savings goals within a specific time period, etc.; the cash flow constraints are clear definitions of the user's restrictions on the time of capital inflow and outflow, flexibility of amount, etc., such as the upper limit of idle funds available for investment each month, the amount of emergency funds to be reserved in the short term, etc.; the preset financial market product database is a resource library containing detailed information on various financial products, such as product characteristics, investment thresholds, risk-return characteristics, etc.; the candidate financial product set is based on the user's cash flow constraints, A portfolio of financial products that may meet user needs is preliminarily screened out from the financial market product database; the risk level identifier is a quantitative assessment of the risk level of the financial product, reflecting the degree of loss that may occur to the product from low to high under the influence of factors such as market fluctuations; the income cycle parameter refers to the time interval and method for the financial product to obtain income, such as daily, monthly, quarterly or annual income settlement and distribution and other related settings. Furthermore, the collection of the financial planning benchmark and cash flow constraints input by the financial user in real time can be achieved through a user interaction interface combined with natural language processing technology. After the user enters relevant information in the interface, the system uses natural language processing technology to convert it into recognizable data; the identification of the risk level identifier and income cycle parameter in the candidate financial product set can be achieved by querying the attribute field of the corresponding product in the financial market product database, and extracting the pre-set risk level identifier field and income cycle parameter field data from the database.

[0032] Specifically, the combination of a preset financial market product database and the financial demand profile to screen a set of candidate financial products that meet the funds flow constraint conditions includes: Extracting an identifier from the financial demand profile to obtain a financial demand identifier; Extracting product feature identifiers from the financial market product database; calculating an identifier matching degree between the financial demand identifier and the product feature identifier, and extracting an initial financial product from the financial market product database based on the identifier matching degree; The initial financial products are subjected to yield adaptation processing to obtain a set of candidate financial products meeting the fund flow constraint condition.

[0033] It should be explained that the financial demand identifier is the identification information extracted from the financial demand portrait that directly reflects the key financial demand characteristics such as the user's occupational attributes, income level, risk preference, etc.; the product feature identifier is the identification content in the financial market product database used to characterize the core attributes such as product type, investment period, risk level, and return level of various financial products; the identifier matching degree represents a quantitative value of the similarity between the financial demand identifier and the product feature identifier, and the higher the value, the higher the matching degree; the initial financial product is a financial product extracted from the financial market product database based on the identifier matching degree, and the identifier matching degree reaches a certain standard and preliminarily meets the user's financial needs.

[0034] Furthermore, the financial demand portrait can be identified and extracted through the attribute extraction algorithm in data mining technology to obtain a financial demand identification; the product feature identification in the financial market product database can be extracted through database query statements and field extraction functions; based on the identification matching degree, initial financial products are extracted from the financial market product database, such as selecting financial products with a matching degree higher than a preset threshold (for example, 0.6) as initial financial products; the initial financial products are subjected to yield adaptation processing to obtain a set of candidate financial products that meet the funds flow constraint conditions, such as based on the amount of funds that can be invested, the expected yield target, etc. in the funds flow constraint conditions, eliminating products with too high or too low yields that do not meet the conditions, and sorting the remaining products according to the balance between yield and risk, and finally determining the set of candidate financial products.

[0035] Furthermore, as an optional embodiment of the present invention, the calculating the identifier matching degree between the financial demand identifier and the product feature identifier includes: Performing vectorization processing on the financial demand identifier and the product feature identifier to obtain a demand identifier vector and a feature identifier vector; Calculate the vector matching degree between the requirement identification vector and the feature identification vector: ; in, Indicates the vector matching degree between the requirement identification vector and the feature identification vector, represents the ath vector in the demand identification vector, represents the bth vector in the feature identification vector, a and b represent the serial numbers corresponding to the requirement identification vector and the feature identification vector respectively, and m and n represent the number corresponding to the requirement identification vector and the feature identification vector respectively; Based on the vector matching degree, an identification matching degree between the financial demand identification and the product feature identification is obtained.

[0036] It should be explained that the demand identification vector and the feature identification vector are respectively formed by quantizing and dimensionally constructing the financial demand identification and the product feature identification, with each identification information being in numerical form as an element of the vector. They are multidimensional vectors that can comprehensively and accurately characterize the financial demand characteristics and financial product characteristics; the vector matching degree represents a quantitative indicator of the degree of similarity between the demand identification vector and the feature identification vector in terms of direction and length ratio, and its numerical value reflects the degree of matching between the financial demand and the financial product characteristics.

[0037] Furthermore, the financial demand identifier and the product feature identifier can be vectorized using the above-mentioned Word2Vecxx model to obtain a demand identifier vector and a feature identifier vector; and the vector matching degree is directly used as the identifier matching degree between the financial demand identifier and the product feature identifier.

[0038] S3. Calculate the matching preference coupling degree between the risk preference label and the risk level identifier, calculate the synergy index between the financial planning benchmark and the revenue cycle parameter, and generate a multidimensional matching index for the candidate financial product set by combining the matching preference coupling degree and the synergy index.

[0039] By calculating the matching preference coupling degree between the risk preference label and the risk level identifier, the present invention can clearly present the degree of fit between the user's subjective risk-bearing willingness and the objective risk level of the financial product, help users quickly identify financial products that match their own risk preferences, reduce investment risks, and improve the rationality of investment decisions. It should be explained that the matching preference coupling degree represents the degree of fit between the risk preference label and the risk level identifier, and is a quantitative indicator for measuring the degree of fit and fit between the user's subjective risk-bearing willingness and the objective risk level of the financial product. The higher the value, the better the match between the two.

[0040] In detail, the calculation of the matching preference coupling degree between the risk preference label and the risk level identifier includes: Encoding the risk preference label and the risk level identifier to obtain a risk preference coding value and a risk level coding value; Normalizing the risk preference code value and the risk level code value to obtain a normalized preference code value and a normalized level code value; In combination with the normalized preference code value and the normalized grade code value, the matching preference coupling degree between the risk preference label and the risk grade identifier is calculated by the following formula: ; in, Indicates the matching preference coupling between the risk preference label and the risk level identifier, represents the normalized preference encoding value, represents the normalized rank-coded value, Indicates selecting the maximum value between the normalized preference coding value and the normalized rank coding value.

[0041] It should be explained that the risk preference coding value and the risk level coding value are respectively numerical values ​​with quantitative meanings obtained after the risk preference label and the risk level identifier are converted according to specific coding rules, and are used to represent risk preference and risk level in digital form; the normalized preference coding value and the normalized level coding value are respectively the risk preference coding value and the risk level coding value after normalization processing, which are mapped to a specific interval for the purpose of comparison, analysis and calculation.

[0042] Furthermore, the risk preference label and the risk level identifier can be encoded separately using a custom sequential coding rule to obtain a risk preference coding value and a risk level coding value. For example, risk preference labels such as conservative, stable, and aggressive are encoded as 1, 2, and 3 in sequence, and risk level identifiers such as low risk, medium-low risk, etc. are encoded as 1 to 5 in sequence; the risk preference coding value and the risk level coding value can be normalized using the maximum-minimum normalization method to obtain a normalized preference coding value and a normalized level coding value, that is, the coding value is mapped to the [0,1] interval, and the formula is (original value-minimum value) / (maximum value-minimum value).

[0043] The present invention calculates a synergy index between the financial planning benchmark and the revenue cycle parameters. Through the synergy index, the compatibility between the user's financial goals and the revenue rhythm of financial products can be understood, thereby helping the user to screen out financial products whose revenue acquisition mode is highly coordinated with their own capital planning needs, improve the efficiency of capital utilization, and realize the orderly advancement of financial planning. It should be explained that the synergy index represents the numerical value of the coordination and fit between the financial planning benchmark and the revenue cycle parameters. The higher the value, the better the synergy between the two.

[0044] In detail, the calculating of the synergy index between the financial planning benchmark and the revenue cycle parameter includes: Performing planning target analysis on the financial planning benchmark to obtain a financial target time domain; Performing segmented quantization processing on the revenue cycle parameter to obtain segmented cycle values; The matching degree between the financial target time domain and the period segment value is calculated to obtain the time matching degree: The matching degree between the financial target time domain and the period segment value is calculated using the following formula: ; Performing weight distribution processing on the time matching degree to obtain a matching degree weight; The synergy index between the financial planning benchmark and the revenue cycle parameter is calculated by combining the matching weight and the time matching.

[0045] It should be explained that the financial target time domain is a specific time range clearly defined in the financial planning benchmark and used to achieve various financial targets (such as asset accumulation, debt repayment, etc.), covering key information such as start time and end time; the cycle segmentation value is the time interval division reflected in the income cycle parameter for the settlement and distribution of financial product income, such as specific segmentation settings by day, month, quarter, year, etc.; the time matching degree represents a quantitative indicator calculated by a specific algorithm (such as a time series comparison algorithm) between the financial target time domain and the cycle segmentation value to measure the degree of fit between the two in terms of time rhythm. The higher the value, the higher the fit; the matching degree weight represents the relative importance of the time matching degree in the comprehensive evaluation of the synergistic relationship between the financial planning benchmark and the income cycle parameter. The higher the weight, the greater the impact of the time matching degree on the overall synergy evaluation.

[0046] Furthermore, the financial planning benchmark can be analyzed for planning objectives through a semantic analysis algorithm, and key time information can be accurately extracted from the complex financial planning description set by the user to obtain the financial target time domain; the revenue cycle parameters can be segmented and quantified by constructing a mathematical model based on the revenue settlement law, and various revenue cycle forms can be converted into unified and comparable numerical segments to obtain cycle segment values, such as the Markov chain model. If the changes in the revenue cycle have certain probabilistic and state transition characteristics, the Markov chain model can be used. Assuming that the revenue cycle has different states, such as a high-return state and a low-return state, and that there is a certain probability of transition between different states, the state transition probability matrix is ​​determined by analyzing historical revenue data, and then the revenue situation in different time periods in the future is predicted based on the current revenue state and transition probability, thereby segmenting and quantifying the revenue cycle. For example, the time period in the same state can be divided into a segment and assigned corresponding numerical characteristics as the cycle segment value; the time matching degree can be weighted by the hierarchical analysis method and with reference to the experience of financial experts and user demand survey data, and the importance of the time matching factor in the overall collaborative evaluation can be comprehensively weighed to obtain the matching weight; combining the matching weight and the time matching degree, the synergy index between the financial planning benchmark and the revenue cycle parameters can be calculated by a weighted sum formula, and the formula is synergy index = time matching degree × matching weight.

[0047] The present invention generates a multidimensional matching index of the candidate financial product set by combining the matching preference coupling degree and the synergy index. The multidimensional matching index can be used to comprehensively consider the risk preference and the degree of financial planning adaptability, thereby providing an important basis for the subsequent generation of the financial product recommendation sequence of the candidate financial product set. It should be explained that the multidimensional matching index represents an indicator of the multi-faceted matching degree of the candidate financial product set. Furthermore, the dimensional importance corresponding to the matching preference coupling degree and the synergy index is analyzed, and the matching preference coupling degree and the synergy index are multiplied by the corresponding dimensional importance respectively to obtain the multidimensional matching index of the candidate financial product set.

[0048] S4. Monitor the rate fluctuation data and credit risk indicators in the financial market of the financial user, calculate the settlement prediction value of the candidate financial product set in combination with the rate fluctuation data and the credit risk indicator, and generate a financial product recommendation sequence for the candidate financial product set in combination with the settlement prediction value and the multidimensional matching index.

[0049] The present invention calculates the settlement forecast value of the candidate financial product set by combining the rate fluctuation data and the credit risk index. The estimated return of each product in the candidate financial product set can be understood through the settlement forecast value, thereby improving the accuracy of generating the financial product recommendation sequence of the candidate financial product set. It should be explained that the rate fluctuation data and the credit risk index are important data information in the financial market of the financial user that can intuitively reflect the dynamic changes in market transaction costs (such as fluctuations in interest rates and handling rates) and the user's own credit reliability (covering aspects such as default possibility and credit rating), and have a key impact on the returns of financial products. The settlement forecast value represents the future return evaluation value of each product in the candidate financial product set. Furthermore, the monitoring of the rate fluctuation data and credit risk indicators in the financial market of the financial user can be achieved through a professional financial data platform.

[0050] In detail, the calculation of the settlement prediction value of the candidate financial product set by combining the rate fluctuation data and the credit risk indicator includes: Extracting a time series rate value from the rate fluctuation data, and smoothing the time series rate value to obtain a smoothed time series rate value; Calculating a rate trend coefficient corresponding to the rate fluctuation data based on the smoothed time-series rate value; Calculating the risk-adjusted value corresponding to the credit risk indicator; The settlement forecast value of the candidate financial product set is calculated by combining the rate trend coefficient and the risk adjustment value using the following formula: ; in, represents the settlement forecast value of the candidate financial product set, Indicates the benchmark settlement value, represents the rate trend coefficient, Represents the risk-adjusted value.

[0051] It should be explained that the time series rate value is the specific rate value corresponding to each time point recorded in chronological order in the rate fluctuation data, which is used to present the original state of the rate in the time series; the smoothed time series rate value is the time series rate value after data smoothing processing, eliminating some random fluctuations, and better reflecting the long-term trend of rate changes; the rate trend coefficient represents the rate fluctuation data calculated through regression analysis, differential operation and other methods, which is used to quantitatively describe the rate and direction of increase or decrease of the rate within a certain period of time, and reflects the degree of its overall change trend. The benchmark settlement value is the historical average settlement value of the financial product.

[0052] Furthermore, the time series rate values ​​in the rate fluctuation data can be extracted in chronological order through a time series data extraction algorithm, and the time series rate values ​​can be smoothed through data smoothing techniques such as the moving average method and the exponential smoothing method to obtain smoothed time series rate values; based on the smoothed time series rate values, the rate trend coefficient corresponding to the rate fluctuation data can be calculated through least squares linear regression fitting; the risk-adjusted value corresponding to the credit risk indicator can be calculated through a risk pricing model and combined with the risk exposure degree of the financial product, such as using the capital asset pricing model (CAPM), based on the systemic risk and market risk premium of the financial product, combined with the specific amount of its risk exposure, to accurately calculate the risk-adjusted value, so as to reasonably reflect the impact of credit risk on product returns.

[0053] The present invention generates a financial product recommendation sequence for the candidate financial product set by combining the settlement prediction value and the multidimensional matching index, and can obtain the recommendation order of the candidate financial product set. It should be explained that the financial product recommendation sequence is the product recommendation order of the candidate financial product set. Furthermore, the settlement prediction value and the multidimensional matching index are combined to generate a financial product recommendation sequence for the candidate financial product set. For example, the settlement prediction value and the multidimensional matching index are first sorted from high to low and assigned corresponding ranking scores, and then the two scores of each product are added together, and the candidate financial product set is sorted according to the total score to generate a financial product recommendation sequence.

[0054] S5. Extract the frequency of money usage and historical redemption operation records from the consumption behavior characteristics, optimize the financial product recommendation sequence based on the frequency of money usage and the historical redemption operation records, and obtain an optimized product recommendation sequence. Based on the optimized product recommendation sequence and the set of candidate financial products, generate a product recommendation set for the financial user.

[0055] The present invention optimizes the financial product recommendation sequence by combining the frequency of money use and the historical redemption operation records, generates a personalized product recommendation set, and improves the practicality of financial product recommendations and user satisfaction. It should be explained that the frequency of money use and the historical redemption operation records are key quantitative indicators in the consumption behavior characteristics, the former reflects the frequency of users using money in various consumption or investment activities, and the latter records in detail the relevant operation details of users' past redemption of financial products; the optimized product recommendation sequence is the financial product recommendation sequence based on the user behavior characteristics reflected by the frequency of money use and the historical redemption operation records, and is generated after adjustment and sorting. It is a recommendation list that can more accurately match the user's actual capital operation habits and financial management operation preferences. Furthermore, the extraction of the frequency of money use and the historical redemption operation records in the consumption behavior characteristics can be achieved through an extraction function, and the extraction function is compiled by a programming language.

[0056] Specifically, the optimizing process of the financial product recommendation sequence based on the frequency of fund usage and the historical redemption operation records to obtain an optimized product recommendation sequence includes: Extracting the redemption operation behavior and the redemption operation object from the historical redemption operation record; Analyzing the coupling mechanism between the redemption operation behavior and the redemption operation object; Sorting the usage frequencies of the funds to obtain sorted usage frequencies of the funds; The financial product recommendation sequence is optimized by combining the coupling mechanism and the usage frequency of the ranked items to obtain an optimized product recommendation sequence.

[0057] It should be explained that the redemption operation behavior and the redemption operation object are respectively the specific behavioral manifestations in the historical redemption operation records (such as redemption time, method, motivation and other behavioral related information) and the target financial products (or assets, etc.); the coupling mechanism is the internal relationship pattern between the redemption operation behavior and the redemption operation object through mutual influence and synergy through specific logic and rules; the sorted frequency of use of funds is the result obtained after sorting the frequency of use of funds (arranged in order from high to low or from low to high to clarify the relative position of the frequency of fund use in different time periods).

[0058] Furthermore, a data parsing algorithm can be used to accurately extract redemption operation behaviors and redemption operation objects from the historical redemption operation records. For example, regular expression matching technology can be used to accurately capture behavioral data such as redemption time and amount from the redemption record text, and the corresponding redemption operation object, that is, the name or code of the financial product involved, can be identified through a data mapping table. An association rule mining algorithm can be used to deeply analyze the coupling mechanism between the redemption operation behavior and the redemption operation object. For example, an Apriori algorithm can be used to mine frequent item sets and strong association rules between behavioral characteristics such as redemption time, redemption amount, and redemption reason and different types of redemption operation objects (such as stock funds, bonds, and time deposits), thereby revealing hidden synergistic relationship patterns between the two. The frequency of use of the funds can be sorted using a sorting function, such as a bubble sort algorithm, to obtain a sorted frequency of use of the funds. Combining the coupling mechanism and the sorted frequency of use of the funds, the financial product recommendation sequence can be optimized to obtain an optimized product recommendation sequence. For example, according to the coupling mechanism, product categories frequently associated with high frequency of use of the funds can be placed in front, and then the order of products of the same category can be adjusted from high to low based on the sorted frequency of use of the funds to reconstruct the recommendation sequence.

[0059] The present invention generates a product recommendation set for the financial user based on the optimized product recommendation sequence and the candidate financial product set, thereby obtaining the best product recommendation for the financial user.

[0060] Compared with the problems described in the background technology, the present invention can gain insight into the user's economic background and professional characteristics by analyzing the occupational attributes and income levels in the user's identity data, and provide a basis for the positioning of financial services; extracting the risk preference labels and consumption behavior characteristics in the historical settlement records can accurately grasp the user's investment tendencies and consumption habits, and provide an important basis for the subsequent construction of the financial demand portrait exclusive to the financial user. Furthermore, the present invention can locate financial products that meet the user's financial status, capital flow needs and risk preferences by combining a preset financial market product database with the financial demand portrait, thereby greatly reducing the time cost of users screening among massive products and improving the accuracy and adaptability of financial services. The present invention calculates the risk preference labels and the risk The matching preference coupling degree between the grade identifiers can clearly show the degree of fit between the user's subjective risk-bearing willingness and the objective risk level of the financial product, helping users to quickly identify financial products that match their own risk preferences, reduce investment risks, and improve the rationality of investment decisions. Furthermore, the present invention calculates the settlement forecast value of the candidate financial product set by combining the rate fluctuation data and the credit risk indicator. The estimated return of each product in the candidate financial product set can be understood through the settlement forecast value, thereby improving the generation accuracy of the financial product recommendation sequence of the candidate financial product set. Furthermore, the present invention optimizes the financial product recommendation sequence by combining the frequency of use of the funds and the historical redemption operation records, generates a personalized product recommendation set, and improves the practicality of financial product recommendations and user satisfaction. Therefore, the intelligent financial product matching method and system provided in the embodiment of the present invention can improve the accuracy of financial product matching.

[0061] Example 2: like Figure 2 The figure shows the functional module diagram of the intelligent financial product matching system of the present invention.

[0062] The intelligent financial product matching system 200 described in the present invention can be installed in an electronic device. Depending on the functionality implemented, the intelligent financial product matching system may include a financial needs profile building module 201, a financial product screening module 202, a multi-dimensional matching index generation module 203, a financial product recommendation sequence generation module 204, and a recommendation sequence optimization module 205. A module, also referred to as a unit, is a series of computer program segments that can be executed by an electronic device processor and perform a fixed function, and is stored in the electronic device's memory.

[0063] In the embodiment of the present invention, the functions of each module / unit are as follows: The financial demand profile building module 201 is used to obtain the user identity data and historical settlement records of the financial user, analyze the occupational attributes and income level in the user identity data, extract the risk preference label and consumption behavior characteristics in the historical settlement records, and combine the occupational attributes, income level and risk preference label to build a financial demand profile unique to the financial user; The financial product screening module 202 is configured to collect the financial planning benchmarks and cash flow constraints input by the financial user in real time, and screen a set of candidate financial products that meet the cash flow constraints based on a preset financial market product database and the financial demand profile, and identify risk level identifiers and return cycle parameters in the set of candidate financial products; The multi-dimensional matching index generating module 203 is configured to calculate the matching preference coupling degree between the risk preference label and the risk level identifier, calculate the synergy index between the financial planning benchmark and the revenue cycle parameter, and generate a multi-dimensional matching index for the candidate financial product set by combining the matching preference coupling degree and the synergy index; The financial product recommendation sequence generation module 204 is configured to monitor rate fluctuation data and credit risk indicators in the financial market of the financial user, calculate settlement prediction values ​​for the candidate financial product set based on the rate fluctuation data and the credit risk indicators, and generate a financial product recommendation sequence for the candidate financial product set based on the settlement prediction values ​​and the multi-dimensional matching index; The recommendation sequence optimization module 205 is used to extract the frequency of payment usage and historical redemption operation records from the consumption behavior characteristics, and optimize the financial product recommendation sequence based on the frequency of payment usage and the historical redemption operation records to obtain an optimized product recommendation sequence. Based on the optimized product recommendation sequence and the set of candidate financial products, a product recommendation set for the financial user is generated.

[0064] In detail, each module in the intelligent financial product matching system 200 in the embodiment of the present invention adopts the same Figure 1 The technical means are the same as the smart financial product matching method described in and can produce the same technical effects, so I will not go into details here.

[0065] It will be apparent to those skilled in the art that the present invention is not limited to the details of the exemplary embodiments described above, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0066] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not limiting. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present invention may be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present invention.

Claims

1. Intelligent financial product matching method, characterized in that: The method comprises: Obtaining the user identity data and historical settlement records of the financial user, parsing the occupational attributes and income level in the user identity data, extracting the risk preference labels and consumption behavior characteristics in the historical settlement records, and combining the occupational attributes, income level, and risk preference labels to construct a unique financial needs profile of the financial user; Collect the financial planning benchmarks and funds flow constraints input by the financial user in real time, combine them with a preset financial market product database and the financial demand profile, screen a set of candidate financial products that meet the funds flow constraints, and identify the risk level identifiers and return cycle parameters in the candidate financial product set; Calculating the matching preference coupling degree between the risk preference label and the risk level identifier, calculating the synergy index between the financial planning benchmark and the revenue cycle parameter, and combining the matching preference coupling degree and the synergy index to generate a multi-dimensional matching index for the candidate financial product set; Monitoring rate fluctuation data and credit risk indicators in the financial market of the financial user, calculating settlement prediction values ​​for the candidate financial product set based on the rate fluctuation data and the credit risk indicators, and generating a financial product recommendation sequence for the candidate financial product set based on the settlement prediction values ​​and the multi-dimensional matching index; Extract the frequency of money use and historical redemption operation records from the consumption behavior characteristics, combine the frequency of money use and the historical redemption operation records, optimize the financial product recommendation sequence, and obtain an optimized product recommendation sequence. Based on the optimized product recommendation sequence and the set of candidate financial products, generate a product recommendation set for the financial user.

2. The intelligent financial product matching method according to claim 1, characterized in that: The extracting of risk preference labels and consumption behavior characteristics from the historical settlement records includes: Cleaning the historical settlement records to obtain target settlement records; extracting record features from the target settlement record, and performing clustering processing on the record features to obtain clustered record features; Analyzing the characteristic behavior tags corresponding to the cluster record features, performing risk scoring processing on the characteristic behavior tags to obtain a risk score value; Determining a risk preference label in the characteristic behavior label based on the risk score value; Based on the characteristic behavior labels, consumption behavior features of the cluster record features are extracted.

3. The intelligent financial product matching method according to claim 1, characterized in that: The step of combining the occupational attributes, the income level, and the risk preference label to construct a financial demand profile exclusive to the financial user includes: Performing standard classification processing on the occupational attributes to obtain standard occupational categories; Performing quantitative grading processing on the income levels to obtain quantitative income grades; Performing weight assignment on the risk preference labels to obtain label weights; fusing the standard occupational category, the income quantification level, and the label weight to obtain a financial demand feature vector of the financial user; determining the financial demand characteristics of the financial user based on the financial demand characteristic vector; The financial demand characteristics are personalized and adapted to obtain a financial demand profile that is exclusive to the financial user.

4. The intelligent financial product matching method according to claim 1, wherein: The screening of a set of candidate financial products that meet the funds flow constraint conditions by combining a preset financial market product database with the financial demand profile includes: Extracting an identifier from the financial demand profile to obtain a financial demand identifier; Extracting product feature identifiers from the financial market product database; calculating an identifier matching degree between the financial demand identifier and the product feature identifier, and extracting an initial financial product from the financial market product database based on the identifier matching degree; The initial financial products are subjected to yield adaptation processing to obtain a set of candidate financial products meeting the fund flow constraint condition.

5. The intelligent financial product matching method according to claim 4, characterized in that: The calculating the identifier matching degree between the financial demand identifier and the product feature identifier includes: Performing vectorization processing on the financial demand identifier and the product feature identifier to obtain a demand identifier vector and a feature identifier vector; Calculate the vector matching degree between the requirement identification vector and the feature identification vector: ； in, Indicates the vector matching degree between the requirement identification vector and the feature identification vector, represents the ath vector in the demand identification vector, represents the bth vector in the feature identification vector, a and b represent the serial numbers corresponding to the requirement identification vector and the feature identification vector respectively, and m and n represent the number corresponding to the requirement identification vector and the feature identification vector respectively; Based on the vector matching degree, an identification matching degree between the financial demand identification and the product feature identification is obtained.

6. The intelligent financial product matching method according to claim 1, characterized in that: The calculating the matching preference coupling degree between the risk preference label and the risk level identifier includes: Encoding the risk preference label and the risk level identifier to obtain a risk preference coding value and a risk level coding value; Normalizing the risk preference code value and the risk level code value to obtain a normalized preference code value and a normalized level code value; In combination with the normalized preference code value and the normalized grade code value, the matching preference coupling degree between the risk preference label and the risk grade identifier is calculated by the following formula: ； in, Indicates the matching preference coupling between the risk preference label and the risk level identifier, represents the normalized preference encoding value, represents the normalized rank-coded value, Indicates selecting the maximum value between the normalized preference coding value and the normalized rank coding value.

7. The intelligent financial product matching method according to claim 1, wherein: The calculating of the synergy index between the financial planning benchmark and the revenue cycle parameter includes: Performing planning target analysis on the financial planning benchmark to obtain a financial target time domain; Performing segmented quantization processing on the revenue cycle parameter to obtain segmented cycle values; The matching degree between the financial target time domain and the period segment value is calculated to obtain the time matching degree: The matching degree between the financial target time domain and the period segment value is calculated using the following formula: ； 1. Performing weight distribution processing on the time matching degree to obtain a matching degree weight; The synergy index between the financial planning benchmark and the revenue cycle parameter is calculated by combining the matching weight and the time matching.

8. The intelligent financial product matching method according to claim 1, wherein: The calculating the settlement prediction value of the candidate financial product set by combining the rate fluctuation data and the credit risk indicator includes: Extracting a time series rate value from the rate fluctuation data, and smoothing the time series rate value to obtain a smoothed time series rate value; Calculating a rate trend coefficient corresponding to the rate fluctuation data based on the smoothed time-series rate value; Calculating the risk-adjusted value corresponding to the credit risk indicator; The settlement forecast value of the candidate financial product set is calculated by combining the rate trend coefficient and the risk adjustment value using the following formula: ； in, represents the settlement forecast value of the candidate financial product set, Indicates the benchmark settlement value, represents the rate trend coefficient, Represents the risk-adjusted value.

9. The intelligent financial product matching method according to claim 1, wherein: The optimizing process of the financial product recommendation sequence based on the frequency of fund usage and the historical redemption operation records to obtain an optimized product recommendation sequence includes: Extracting the redemption operation behavior and the redemption operation object from the historical redemption operation record; Analyzing the coupling mechanism between the redemption operation behavior and the redemption operation object; Sorting the usage frequencies of the funds to obtain sorted usage frequencies of the funds; The financial product recommendation sequence is optimized by combining the coupling mechanism and the usage frequency of the ranked items to obtain an optimized product recommendation sequence.

10. Intelligent financial product matching system, characterized by: The system comprises: A financial needs profile building module is used to obtain the user identity data and historical settlement records of a financial user, analyze the occupational attributes and income level in the user identity data, extract the risk preference labels and consumption behavior characteristics in the historical settlement records, and combine the occupational attributes, income level and risk preference labels to build a financial needs profile unique to the financial user; A financial product screening module is configured to collect the financial planning benchmarks and cash flow constraints input by the financial user in real time, and screen a set of candidate financial products that meet the cash flow constraints based on a preset financial market product database and the financial demand profile, and identify the risk level identifiers and return cycle parameters in the set of candidate financial products; a multidimensional matching index generation module, configured to calculate a matching preference coupling degree between the risk preference label and the risk level identifier, calculate a synergy index between the financial planning benchmark and the revenue cycle parameter, and generate a multidimensional matching index for the candidate financial product set by combining the matching preference coupling degree and the synergy index; a financial product recommendation sequence generation module, configured to monitor rate fluctuation data and credit risk indicators in the financial market of the financial user, calculate settlement prediction values ​​for the candidate financial product set based on the rate fluctuation data and the credit risk indicators, and generate a financial product recommendation sequence for the candidate financial product set based on the settlement prediction values ​​and the multi-dimensional matching index; The recommendation sequence optimization module is used to extract the frequency of payment usage and historical redemption operation records from the consumption behavior characteristics, and optimize the financial product recommendation sequence based on the frequency of payment usage and the historical redemption operation records to obtain an optimized product recommendation sequence. Based on the optimized product recommendation sequence and the set of candidate financial products, a product recommendation set for the financial user is generated.

Images