Automobile financial risk intelligent identification method and system

Through the dynamic adjustment of the model, we can monitor the changes in market and borrower credit data in real time, simulate future market conditions and behavior changes, solve the problems of lagging and inaccurate risk assessment in the existing technology, and realize the efficient identification and evaluation of automobile financial risks , reduces the bad debt rate and improves decision-making efficiency.

CN119005991AInactive Publication Date: 2024-11-22CHENGDU WANWANG SECONDARY PLANET COMM EQUIP CO LTD
View PDF 0 Cites 7 Cited by

Patent Information

Application Number
CN202411489195.0
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-10-24
Publication Date
2024-11-22
Estimated Expiration
Not applicable · inactive patent

AI Technical Summary

Technical Problem

The existing intelligent identification methods for automobile finance risks rely on static data analysis and lack immediate response capabilities to market changes, resulting in lagging or inaccurate risk assessment results, unable to quickly adjust the model, and increasing potential losses of financial institutions.

Method used

By collecting market interest rates, borrower credit records and loan terms data, a behavioral market relationship framework is formed, the changes in market and borrower credit data are tracked in real time, the connection weights of internal variables of the model are dynamically adjusted, the market dynamic interaction model is constructed, and future market conditions are simulated. Borrower behavior changes, regularly update default probability and credit risk, optimize risk levels, and extract keywords based on text data analysis to establish an automobile financial risk identification model.

Benefits of technology

It has achieved efficient adaptability and response speed to automobile financial risks, accurately captured the related information between market interest rates, borrower credit records and loan terms, provided efficient risk assessment tools, reduced bad debt rates, improved customer service experience and Decision-making efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119005991A_ABST
    Figure CN119005991A_ABST
Patent Text Reader

Abstract

The invention relates to the technical field of financial science and technology, in particular to an automobile financial risk intelligent identification method and system, and the method comprises the following steps: collecting the market interest rate, borrower credit record and loan term data based on the demands of an automobile financial market, analyzing the relation between the data, and forming a behavior market relation framework. According to the method, the adaptability and response speed to automobile financial risks are improved by monitoring changes of the market data and the borrower credit data in real time and dynamically adjusting the connection weight between the internal variables of the model, and the association among the market interest rate, the borrower credit record and the loan terms is accurately captured by comparing data at different time points. Understanding of market and borrower behavior interaction modes is increased, changes of future market and borrower behaviors are simulated, an efficient risk assessment tool is provided, text data analysis is refined, risk index capture is enhanced through keyword extraction, credit and loan resource allocation is optimized, the bad debt rate is reduced, and decision-making efficiency is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of financial technology, and in particular to a method and system for intelligently identifying automobile financial risks. Background Art

[0002] The field of financial technology involves the use of modern technology to improve and automate financial services and management processes. This field combines finance, computer science and information technology, aiming to improve the efficiency of financial institutions' services, reduce costs, enhance security and provide personalized financial solutions. Fintech includes mobile payments, automated investment management (such as robot advisors), smart contracts, artificial intelligence tools for creditor risk assessment, and cryptocurrency trading platforms. This technology field continues to promote innovation in financial services, making traditional banking and emerging financial services such as crowdfunding more convenient and intelligent.

[0003] Among them, the intelligent identification method of automobile financial risks is a method that uses artificial intelligence technology to identify and evaluate credit risks in the field of automobile finance. This topic involves the development and implementation of algorithms or models. By analyzing multi-dimensional information such as borrowers' financial data, credit history, car purchase intentions and related market economic factors, it can predict and prevent potential default risks. The main purpose of this method is to help financial institutions reduce bad debt rates, optimize credit resource allocation, and improve customer service experience and decision-making efficiency, thereby gaining an advantage in the highly competitive market.

[0004] Although existing technologies in the field of financial technology have integrated a variety of advanced technologies, such as robot advisors and smart contracts, there are still some limitations in the intelligent identification of automotive financial risks. Existing methods mostly rely on static data analysis and lack the ability to respond immediately to market changes. This results in financial institutions being unable to quickly adjust risk assessment models when faced with drastic market changes, which can lead to delayed or inaccurate risk assessment results. For example, when there is an emergency in the credit market, traditional models cannot immediately reflect the borrower's true risk status, increasing the potential losses of financial institutions. When processing unstructured text data in loan applications, existing technologies do not fully utilize risk-sensitive information in the data and fail to effectively identify key risk indicators, which to a certain extent limits the accuracy and efficiency of risk management. Summary of the invention

[0005] The purpose of the present invention is to solve the shortcomings of the prior art and to propose an intelligent identification method and system for automobile financial risks.

[0006] In order to achieve the above object, the present invention adopts the following technical solution, a method for intelligently identifying automobile financial risks, comprising the following steps: S1: Based on the needs of the auto finance market, collect market interest rates, borrower credit records and loan terms data, analyze the relationship between the data, and form a behavioral market relationship framework; S2: Based on the key variables in the behavioral market relationship framework, real-time tracking of changes in market data and borrower credit data is performed, and the connection weights of internal variables of the model are dynamically adjusted to create a market dynamic interaction model; S3: Using the market dynamic interaction model, simulate the market conditions and borrower behavior changes in the future time period, adjust the model parameters and match the new market conditions, and build a future behavior prediction adjustment model; S4: by adjusting the future behavior prediction model, regularly updating the borrower's default probability and credit risk, adjusting the risk level, and generating risk assessment results; S5: Analyze the text data in the loan application according to the risk assessment results, extract keywords and phrases associated with financial risks, refine the extraction process of key information points, and form a text risk indicator set; S6: combining the data in the text risk indicator set with the risk assessment results, analyzing the matching degree between the keywords and the current risk model, optimizing the correlation analysis between the keywords and the risk level, and establishing an automobile financial risk identification model.

[0007] As a further solution of the present invention, the behavioral market relationship framework includes interaction pattern identification, market behavior response, and periodic changes; the market dynamic interaction model includes weight adjustment indicators, monitoring cycle settings, and data feedback mechanisms; the future behavior prediction adjustment model includes simulation result accuracy, parameter adaptability, and market contingency strategies; the risk assessment results include credit rating updates, default probability corrections, and risk response strategies; the text risk indicator set includes risk semantic analysis, keyword correlation, and information sensitivity; the auto finance risk identification model includes risk level redefinition, keyword utility testing, and model adaptability enhancement.

[0008] As a further solution of the present invention, based on the needs of the automobile finance market, the steps of collecting market interest rates, borrower credit records and loan terms data, analyzing the relationship between the data, and forming a behavioral market relationship framework are as follows: S101: Based on the needs of the auto finance market, collect the original data of market interest rates, borrower credit records and loan terms, classify and mark the data, extract key fields, unify the structure of the data set, adjust the data fields, and establish a unified data set; S102: Based on the unified data set, mark the data records of the differentiated time points, perform data segmentation, associate the time series with the borrower behavior data, analyze the interaction between the borrower behavior and the market conditions through association, and construct a time difference data relationship diagram; S103: Using the time difference data relationship diagram, analyzing the interaction pattern of the differentiated time point data, comparing the data relationships, identifying the behavioral market relationship of the borrower's interaction with the market conditions, and obtaining a behavioral market relationship framework.

[0009] As a further solution of the present invention, according to the key variables in the behavioral market relationship framework, the changes of market data and borrower credit data are tracked in real time, and the connection weights of the internal variables of the model are dynamically adjusted to create a market dynamic interaction model. Specifically, the steps are as follows: S201: According to the behavioral market relationship framework, market dynamic data and borrower credit information are collected in real time, including stock market fluctuation data, commodity trading volume and updates of borrowers' current credit ratings, and synchronized to a central database through a network interface to obtain market credit data; S202: Based on the market credit data, dynamically adjust the connection weights between the internal variables of the model, refer to the frequency and amplitude of data changes and current data input, update the weights in real time through the internal adjustment mechanism, and generate a weight adjustment model; S203: Based on the weight adjustment model, a machine learning algorithm is used to simulate and predict the market dynamics of the interaction between market behavior and the borrower's credit status, collect and organize market behavior data, evaluate the borrower's credit status, perform model verification and adjustment, and obtain a market dynamic interaction model.

[0010] As a further solution of the present invention, the formula of the machine learning algorithm is as follows:

[0011] in, represents the predicted output of the model, is the activation function, is the weight matrix, is the input feature vector, is the bias term, is the adjustment factor, The degree of market competition, is the risk adjustment factor, Score industry risk.

[0012] As a further solution of the present invention, the market dynamic interaction model is used to simulate the market conditions and borrower behavior changes in the future time period, adjust the model parameters and match the new market conditions, and the steps of constructing the future behavior prediction adjustment model are specifically as follows: S301: using the market dynamic interaction model, simulating changes in market conditions and borrower behavior in a future time period, recording key indicators in the output, including interest rate fluctuations and changes in borrowing intentions, marking the indicators with time, classifying trends of differentiated market behaviors, and generating market forecast data records; S302: Based on the market forecast data record, identify the change points in the parameters, adjust the influencing parameters in the model, including adjusting the credit score weight, interest rate response sensitivity, recalibrating for new market data, matching the current changes in the market, and obtaining optimized parameter configuration; S303: According to the optimization parameter configuration, the logical structure of the behavior prediction model in the future time period is updated, the prediction logic is adjusted and the matching ability of the model to the market changes in the future time period is optimized, the interaction mode between the parameters is refined, and the future behavior prediction adjustment model is constructed.

[0013] As a further solution of the present invention, the steps of regularly updating the borrower's default probability and credit risk, adjusting the risk level, and generating the risk assessment result through the future behavior prediction adjustment model are specifically as follows: S401: extracting the default probability and credit risk data of each borrower from the future behavior prediction adjustment model, including account behavior analysis and payment record update, and automatically synchronizing the data from the market monitoring configuration to the risk assessment database every day to obtain a default probability data set; S402: Based on the default probability data set, analyze the current market data changes, adjust the risk level of the borrower item by item, compare the current market dynamics with the historical credit behavior, automatically adjust the risk level threshold, and generate an updated risk level; S403: Use the updated risk level in combination with current market data to conduct risk assessment, analyze the data of each borrower and conduct credit assessment to obtain a risk assessment result.

[0014] As a further solution of the present invention, according to the risk assessment results, the text data in the loan application is analyzed, keywords and phrases associated with financial risks are extracted, the extraction process of key information points is refined, and the steps of forming a text risk indicator set are specifically as follows: S501: Based on the risk assessment results, select words and phrases related to financial risks from the loan application text, identify the words through text tagging technology, classify and organize financial terms, risk descriptions and related indicators, and build a risk vocabulary library; S502: According to the risk vocabulary library, adjust the word frequency analysis parameters, strengthen the context association analysis, optimize the word risk weight, refine the risk indicator extraction, and generate a risk information set; S503: Integrate key risk indicators through the risk information set, collect and analyze related risk data, screen and determine key risk indicators, and generate a text risk indicator set.

[0015] As a further solution of the present invention, the steps of combining the data in the text risk index set with the risk assessment results, analyzing the matching degree between the keywords and the current risk model, optimizing the association analysis between the keywords and the risk level, and establishing the automobile financial risk identification model are as follows: S601: According to the text risk indicator set, data is collected, risk indicator data related to automobile finance is integrated, keywords in the text data are classified and relevance is marked, the marked keywords are used to perform key verification on the risk assessment model, and data matching and preliminary screening are performed to generate preliminary keyword mapping data; S602: Quantitatively analyze the relationship between keywords and risk levels using the preliminary keyword mapping data, remap the risk weights of keywords by statistically analyzing the frequency and correlation, including redefining the connection rules between keywords and risk categories, adjusting output parameters and matching risk rating requirements, and constructing a keyword risk association mapping table; S603: Based on the keyword risk association mapping table, the interaction between risk indicators and keywords is optimized by adjusting model parameters, including adjusting the model structure and updating the risk weights, to obtain an automobile financial risk identification model.

[0016] An intelligent automobile financial risk identification system is provided, and the intelligent automobile financial risk identification system is used to execute the intelligent automobile financial risk identification method. The system comprises: The data capture module collects market interest rates, borrower credit records, and loan terms data based on the needs of the auto finance market, identifies key influencing factors through association analysis, and obtains a data relationship set between the market and borrowers; The market behavior analysis module uses the data relationship set between the market and the borrower to analyze the correlation between the borrower and the market conditions, adopts change point detection, and generates a behavioral market interaction framework; The dynamic interaction model module is based on the behavioral market interaction framework, monitors the changes in market and borrower data in real time, dynamically adjusts the weights in the relationship model, matches market changes, and builds a market dynamic response model; The risk prediction module applies the market dynamic response model to simulate the changes in market and borrower behavior in the future time period, adjusts the risk level according to the model output, adjusts the risk assessment parameters, and obtains the future behavior and risk prediction model; The risk identification optimization module analyzes the text data of the loan application based on the future behavior and risk prediction model, extracts key information points associated with risks, analyzes the information points, optimizes risk ratings, and establishes an automobile financial risk identification model.

[0017] Compared with the prior art, the advantages and positive effects of the present invention are: In the present invention, by real-time monitoring of changes in market data and borrower credit data, the connection weights between internal variables of the model are dynamically adjusted, effectively improving the adaptability and response speed to automobile financial risks. By comparing data analysis at different time points, the correlation information between market interest rates, borrower credit records and loan terms can be captured more accurately, forming a behavioral market relationship framework, which not only increases the understanding of the interaction pattern between borrower behavior and market conditions, but also simulates future changes in market conditions and borrower behavior, providing financial institutions with an efficient risk assessment tool. Through this dynamic adjustment mechanism, financial institutions can more flexibly adjust risk levels and optimize credit resource allocation, thereby reducing bad debt rates, improving customer service experience and decision-making efficiency, and also refining text data analysis, by extracting keywords and phrases related to financial risks, enhancing the ability to capture risk indicators, making risk prediction more accurate and timely. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a schematic diagram of the workflow of the present invention; Figure 2 This is a detailed flow chart of S1 of the present invention; Figure 3 This is a detailed flow chart of S2 of the present invention; Figure 4 This is a detailed flow chart of S3 of the present invention; Figure 5 This is a detailed flow chart of S4 of the present invention; Figure 6 This is a detailed flow chart of S5 of the present invention; Figure 7 This is a detailed flow chart of S6 of the present invention; Figure 8 It is a system flow chart of the present invention. DETAILED DESCRIPTION

[0019] In order to make the purpose, technical solution and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. 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] In the description of the present invention, it should be understood that the terms "length", "width", "up", "down", "front", "back", "left", "right", "vertical", "horizontal", "top", "bottom", "inside" and "outside" etc., indicating directions or positional relationships, are based on the directions or positional relationships shown in the accompanying drawings, and are only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore cannot be understood as a limitation on the present invention. In the description of the present invention, "multiple" means two or more, unless otherwise clearly and specifically defined.

[0021] See also Figure 1 The present invention provides a technical solution, a method for intelligently identifying automobile financial risks, comprising the following steps: S1: Based on the needs of the auto finance market, capture the correlation information of market interest rates, borrower credit records and loan terms, analyze the information and identify the interaction pattern between borrower behavior and market conditions, and form a behavioral market relationship framework by comparing data at different time points; S2: Based on the key variables in the behavioral market relationship framework, the connection weights between the internal variables of the model are dynamically adjusted by real-time monitoring of changes in market data and borrower credit data to create a market dynamic interaction model; S3: Use the market dynamic interaction model to simulate the market conditions and borrower behavior changes in the future time period, adjust the parameters according to the model output and match the new market conditions, and build a future behavior prediction adjustment model; S4: Adjust the model through future behavior prediction, regularly update the borrower's default probability and credit risk, adjust the risk level according to current market data and prediction results, and generate risk assessment results; S5: Based on the risk assessment results, analyze the text data in the loan application, extract keywords and phrases related to financial risks, use text analysis tools to refine the extraction process of key information points, capture the core information related to risks, and form a set of text risk indicators; S6: Combine the data in the text risk indicator set with the risk assessment results, analyze the matching degree between the keywords and the current risk model, optimize the correlation analysis between the keywords and the risk level, adjust the risk rating, and establish an automobile financial risk identification model.

[0022] The behavioral market relationship framework includes interaction pattern identification, market behavior response, and cyclical changes. The market dynamic interaction model includes weight adjustment indicators, monitoring cycle settings, and data feedback mechanisms. The future behavior prediction adjustment model includes simulation result accuracy, parameter adaptability, and market contingency strategies. Risk assessment results include credit rating updates, default probability corrections, and risk response strategies. The text risk indicator set includes risk semantic analysis, keyword relevance, and information sensitivity. The auto finance risk identification model includes risk level redefinition, keyword utility testing, and model adaptability enhancement.

[0023] Specifically, if Figure 2 As shown in the figure, based on the needs of the auto finance market, the steps of collecting market interest rates, borrower credit records and loan terms data, analyzing the relationship between the data, and forming a behavioral market relationship framework are as follows: S101: Based on the needs of the automobile finance market, collect the original data of market interest rates, borrower credit records and loan terms, classify and mark the data, extract key fields, unify the structure of the data set, adjust the data fields, and establish the unified data set. The execution process is as follows; Based on the needs of the auto finance market, we collect raw data on market interest rates, borrower credit records, and loan terms, conduct detailed checks on the collected data to ensure data integrity and accuracy, use data preprocessing techniques such as removing outliers and filling missing values ​​to ensure data quality, classify and label the data, use predefined classification rules, such as classifying borrower credit records into four categories: "excellent", "good", "average", and "poor", and classify them according to the term and interest rate of the loan terms, extract key fields such as loan amount, repayment status, and interest rate, unify the structure of the data set through key fields, adjust the format and unit of the data fields, and establish a unified data set. We optimize the data fields based on actual application needs, such as adjusting the timestamp format, currency unit, and interest rate representation, build a unified data set, and standardize data values The calculation formula is as follows:

[0024] in, represents the original data value, Represents the average value of all corresponding data in the data set. Represents standard deviation.

[0025] S102: Based on the unified data set, the data records of the differentiated time points are marked, data segmentation is performed, the time series is associated with the borrower behavior data, the interaction between the borrower behavior and the market conditions is analyzed through association, and the execution process of constructing the time difference data relationship diagram is as follows; Based on a unified data set, data records at differentiated time points are annotated. This involves defining a time point as a reference, such as the day a loan is issued or the day a loan is repaid. Data segmentation is performed, and the data is divided according to the time series, in units of months or years, and differentiated according to the borrower's repayment behavior and changes in market interest rates. The time series is associated with the borrower's behavior data. The step is performed by calculating the correlation coefficient between each time point in the time series and the borrower's repayment behavior. The interaction between the borrower's behavior and market conditions is analyzed through association, and a time-difference data relationship graph is constructed. Using graph theory technology, each node is defined as a time point, and the edge represents the strength of the relationship between the nodes. The graph constructed in this way can intuitively show the dynamic changes in the relationship between the borrower's behavior and market conditions, and the strength of the relationship. The calculation formula is as follows:

[0026] in, and Represents the borrower at a certain point in time and behavioral variables, such as repayment amount and overdue record.

[0027] S103: Using the time difference data relationship diagram, analyze the interaction pattern of the differentiated time point data, compare the data relationship, identify the behavioral market relationship of the borrower and the market conditions, and obtain the execution process of the behavioral market relationship framework as follows; Use time difference data relationship diagrams to analyze the interaction patterns of data at different time points, including identifying which time points have the most frequent interactions and how interactions affect borrowers' behavior. Compare data relationships and use statistical methods, such as variance analysis, to determine the significant differences in interactions between different time points. Identify the behavioral market relationship of borrowers' interactions with market conditions. By comparing each borrower's behavioral data with market condition data, find out the main market factors that affect borrowers' behavior and obtain a behavioral market relationship framework. This framework is used to predict future borrowers' behavioral trends and provide data support for strategy adjustments, impact assessment The calculation formula is as follows:

[0028] in, The weight coefficient representing the impact of different market conditions on borrower behavior, They represent market interest rates, borrower credit ratings, and economic indicators, respectively.

[0029] Specifically, if Figure 3As shown in the figure, according to the key variables in the behavioral market relationship framework, the changes in market data and borrower credit data are tracked in real time, and the connection weights of the internal variables of the model are dynamically adjusted. The specific steps to create a market dynamic interaction model are as follows: S201: According to the behavioral market relationship framework, market dynamic data and borrower credit information are collected in real time, including stock market fluctuation data, commodity trading volume and updates of borrowers' current credit ratings, and synchronized to the central database through the network interface. The execution process of obtaining market credit data is as follows; According to the behavioral market relationship framework, market dynamic data and borrower credit information are collected in real time, including connecting multiple data sources, such as stock exchanges, credit rating agencies, etc., to ensure the real-time and relevance of data, including stock market fluctuation data, commodity trading volume and updates of borrowers' current credit ratings. Through specially designed data crawling programs, the required data is extracted from various interfaces and synchronized to the central database through the network interface. This involves the use of secure data transmission protocols and data encryption technology to ensure the security and integrity of the data, obtain market credit data, perform preliminary data cleaning and format standardization on the data, ensure that the data can be correctly parsed and used by subsequent processing processes, and improve the efficiency of market credit data synchronization. The calculation formula is as follows:

[0030] in, Represents the time required for data collection and receipt, Represents the data transmission time, Represents data processing and storage time.

[0031] S202: Based on the market credit data, dynamically adjust the connection weights between the internal variables of the model, refer to the frequency and amplitude of data changes and the current data input, and update the weights in real time through the internal adjustment mechanism to generate the execution process of the weight adjustment model as follows; Based on market credit data, the connection weights between variables within the model are dynamically adjusted, including identifying those variables that have the greatest impact on forecast accuracy, such as stock market fluctuations and borrower credit changes. Referring to the frequency and magnitude of data changes and current data inputs, a set of pre-set weight adjustment rules are used. The rules are based on historical data analysis to reflect the latest changes in market and borrower behavior in real time. The weights are updated in real time through an internal adjustment mechanism, including an automatic adjustment algorithm that adjusts the weights according to real-time data liquidity and change rate to maintain the sensitivity and accuracy of the model, and generates a weight adjustment model that can adapt to market changes and adjust its internal structure in real time to better predict future market and credit behavior. The weight adjustment factor The calculation formula is as follows:

[0032] in, Representative variables The latest change in represent The historical volatility standard deviation, represents the learning rate, Represents the old weight.

[0033] S203: Based on the weight adjustment model, a machine learning algorithm is used to simulate and predict the market dynamics of the interaction between market behavior and the borrower's credit status, collect and organize market behavior data, evaluate the borrower's credit status, and perform model verification and adjustment. The execution process of the market dynamic interaction model is as follows; Based on the weight adjustment model, a machine learning algorithm is used to set the initial model parameters according to the current market and borrower behavior data, simulate and predict the market dynamics of the interaction between market behavior and borrower credit status, adjust the model to adapt to the new market environment through in-depth learning and pattern recognition of historical data, collect and organize market behavior data, including regularly extracting data from the market monitoring system and performing necessary data preprocessing, such as normalization and outlier processing, assessing the borrower's credit status, and performing model verification and adjustment. According to the comparison results between the model output and the actual market behavior, the model is slightly adjusted and optimized to obtain a market dynamic interaction model. The model continuously learns through actual data and optimizes its prediction ability to provide more accurate market dynamic predictions.

[0034] The formula for the machine learning algorithm is as follows:

[0035] in, represents the predicted output of the model, is the activation function, is the weight matrix, is the input feature vector, is the bias term, is the adjustment factor, The degree of market competition, is the risk adjustment factor, Score industry risk.

[0036] The execution process is as follows: Determined through market data analysis The value of and industry risk scores , to reflect the specific conditions of different markets and industries, using adjustment coefficients and To refine and The two coefficients are obtained through data regression analysis to ensure that they can effectively reflect the impact of and The real impact in different market environments, all parameters are input into the model, through the activation function After processing, the prediction results are obtained , As a bias term, it adjusts the overall prediction threshold to optimize the prediction accuracy of the model.

[0037] Specifically, if Figure 4 As shown in the figure, the market dynamic interaction model is used to simulate the market conditions and borrower behavior changes in the future time period, adjust the model parameters and match the new market conditions, and the steps to build the future behavior prediction adjustment model are as follows: S301: Use the market dynamic interaction model to simulate the market conditions and borrower behavior changes in the future time period, record the key indicators in the output, including interest rate fluctuations and changes in borrowing intentions, time-mark the indicators, classify the trends of differentiated market behaviors, and generate the market forecast data record. The execution process is as follows; Use the market dynamics interaction model to simulate changes in market conditions and borrower behavior in future time periods. This involves setting the model's forecast range and time resolution, such as predicting daily market dynamics in the next week, recording key indicators in the output, including extracting key data through the model output, such as interest rate fluctuations and daily changes in borrowing intentions, and time-marking the indicators. The steps use timestamps to mark the specific recording time of each data point to ensure the timing accuracy of the data, classify the trends of differentiated market behaviors, and use statistical analysis methods such as trend line analysis and moving averages to identify and classify the main changing trends of market behaviors. Generate market forecast data records to provide support for decision making and market behavior trends. The calculation formula is as follows:

[0038] in, represents the change in interest rate, represents the average interest rate, represents the change in borrowing intention, represents the average borrowing intention, is the adjustment factor.

[0039] S302: Based on the market forecast data records, identify the change points in the parameters, adjust the influencing parameters in the model, including adjusting the credit score weight, interest rate response sensitivity, recalibrating for new market data, matching the current changes in the market, and obtaining the execution process of optimizing the parameter configuration is as follows; Based on the market forecast data records, identify the change points in the parameters, including analyzing the significant data fluctuation points and outliers in the records to determine the model parameters that need to be adjusted, and adjust the influencing parameters in the model. This involves using the data analysis results to update the key parameters in the model, such as credit score weights and interest rate response sensitivity, to reflect the latest market conditions, recalibrate for new market data, and adjust the model through simulation tests and historical data comparison to ensure that its output is consistent with actual market performance and match the current changes in the market. Through continuous model optimization and adjustment, ensure that the model can accurately predict future market conditions, obtain optimized parameter configuration, and adjust parameters. The calculation formula is as follows:

[0040] in, represents the old credit score weight, and Represent the new and old data points respectively, and is the adjustment factor.

[0041] S303: According to the optimized parameter configuration, the logical structure of the behavior prediction model in the future time period is updated, the prediction logic is adjusted and the matching ability of the model to the market changes in the future time period is optimized, the interaction mode between the parameters is refined, and the execution process of constructing the future behavior prediction adjustment model is as follows; According to the optimized parameter configuration, the logical structure of the behavior prediction model in the future time period is updated. This involves redesigning the internal logic and data flow of the model according to the latest market data and forecasting needs, adjusting the forecasting logic and optimizing the model's ability to match market changes in the future time period. Through algorithm adjustment and logic optimization, the model's response speed and accuracy to future market changes are improved, the interaction between parameters is refined, and the adaptability and flexibility of the model are increased by enhancing the linkage and interaction between the internal parameters of the model. By building a future behavior prediction adjustment model, the model will be able to more effectively predict changes in the market and borrower behavior, provide more accurate data support for decision-making, and the model's adaptability. The calculation formula is as follows:

[0042] in, represents the model prediction value, Represents actual market data, is the adaptability coefficient, Represents the number of data points.

[0043] Specifically, if Figure 5As shown in the figure, the steps of adjusting the model through future behavior prediction, regularly updating the borrower's default probability and credit risk, adjusting the risk level, and generating risk assessment results are as follows: S401: Extract the default probability and credit risk data of each borrower from the future behavior prediction adjustment model, including account behavior analysis and payment record update. The data is automatically synchronized from the market monitoring configuration to the risk assessment database every day. The execution process of obtaining the default probability data set is as follows; Extract the default probability and credit risk data of each borrower from the future behavior prediction adjustment model, and extract detailed information about the borrower's account behavior and payment record through the model's data interface. The data covers the key indicators of the borrower's financial behavior and repayment ability, including account behavior analysis and payment record updates. Use data processing algorithms to update each record to ensure that all data is up to date and reflects the borrower's current credit status. The data is automatically synchronized from the market monitoring configuration to the risk assessment database every day. Through scheduled tasks and data synchronization technology, ensure that every piece of data stored in the database is updated in real time, so that the default probability assessment is more accurate, and obtain the default probability data set. This data set is used for further analysis and decision making. The default probability The calculation formula is as follows:

[0044] in, Represents data points extracted from borrower account behavior and payment records, represents the corresponding weight coefficient, represents the total number of data points, is the normalization function.

[0045] S402: Based on the default probability data set, analyze the current market data changes, adjust the risk level of the borrower item by item, compare the current market dynamics with the historical credit behavior, automatically adjust the risk level threshold, and generate the updated risk level. The execution process is as follows; Based on the default probability data set, analyze the current market data changes, identify changes and trends in market conditions through statistical analysis techniques, such as interest rate changes, fluctuations in economic indicators, etc., adjust the risk level of borrowers item by item, and use the predetermined risk rating model to adjust the risk level of each borrower according to the real-time changes in the default probability data and market conditions. Compare the current market dynamics with historical credit behavior, and use data comparison and analysis technology to automatically identify and adjust the risk level of borrowers that have changed significantly, automatically adjust the risk level threshold, and optimize the threshold setting through machine learning algorithms to respond to market changes, generate updated risk levels, and adjust risk levels. The calculation formula is as follows:

[0046] in, Represents the change in credit behavior, represents the historical average credit behavior, Represents the change in market data Represents the historical average of market data, is the adjustment factor.

[0047] S403: Using the updated risk level and combining it with the current market data, a risk assessment is performed, and the data of each borrower is analyzed and passed through a credit assessment. The execution process of obtaining the risk assessment result is as follows; Using the updated risk rating, combined with current market data, through the analysis of the latest market data, such as interest rates, stock market performance, etc., determine the key market variables that affect credit assessment, conduct risk assessment, conduct detailed analysis of each borrower's data, and use the credit assessment model to assess the risk status of each borrower. Through credit assessment, using data analysis and machine learning technology, the borrower's historical credit behavior and current market data are combined to calculate the credit risk of each borrower and obtain the risk assessment results. The results are used for credit management and decision support, providing lending institutions with a data-driven decision-making basis for credit risk assessment. The calculation formula is as follows:

[0048] in, represents the key variables extracted from market data, represents the corresponding influence coefficient, represents the total number of key variables, is the risk assessment factor.

[0049] Specifically, if Figure 6 As shown in the figure, according to the risk assessment results, the text data in the loan application is analyzed, the keywords and phrases associated with financial risks are extracted, the extraction process of key information points is refined, and the steps of forming a text risk indicator set are as follows: S501: Based on the risk assessment results, words and phrases related to financial risks are screened from the loan application text, words are identified through text tagging technology, financial terms, risk descriptions and related indicators are classified and sorted, and the execution process of building a risk vocabulary library is as follows; Based on the risk assessment results, we screen words and phrases related to financial risks from the loan application texts. We use natural language processing techniques, such as keyword extraction and semantic analysis, to identify terms related to financial instability and default risk in the texts. We use text tagging techniques to identify words, which involves using text analysis tools to highlight and tag words for subsequent processing. We classify and organize financial terms, risk descriptions, and related indicators. We use classification algorithms to systematically classify the extracted words according to risk types and financial functions. We build a risk vocabulary library, which will serve as a basic resource for analyzing financial documents and identifying potential risks. We also use vocabulary risk rating. The calculation formula is as follows:

[0050] in, Represents the frequency of occurrence of a word, represents the weight coefficient associated with risk, Represents the total number of words filtered out.

[0051] S502: According to the risk vocabulary library, adjust the word frequency analysis parameters, strengthen the context association analysis, optimize the word risk weight, refine the risk indicator extraction, and generate the risk information set. The execution process is as follows; According to the risk vocabulary library, adjust the frequency analysis parameters, set specific thresholds and parameters to improve the sensitivity and accuracy of the analysis, ensure that risk-related words and phrases can be accurately captured, strengthen contextual association analysis, and use contextual analysis tools to understand the meaning and use of words in specific financial contexts, which will help improve the accuracy of vocabulary classification and optimize vocabulary risk weights. Through a data-driven approach, adjust the risk weight of each word based on historical data and market feedback, refine risk indicator extraction, extract more detailed risk indicators through advanced analysis technology, and generate risk information sets, which will be used for further risk assessment and monitoring. Vocabulary weight adjustment The calculation formula is as follows:

[0052] in, represents the frequency of words, represents the contextual relevance score, and is the adjustment factor.

[0053] S503: The execution process of integrating key risk indicators, collecting and analyzing related risk data, screening and determining key risk indicators, and generating a text risk indicator set through the risk information set is as follows; Through the risk information set, key risk indicators are integrated, including identifying the most critical data points from the risk information set, such as high-risk words and frequently occurring risk patterns, collecting and analyzing related risk data, extracting valuable risk information from large amounts of text data through data mining techniques, such as association rule mining and pattern recognition, screening and determining key risk indicators, determining which indicators are most critical for risk assessment through statistical analysis and expert review, and generating a text risk indicator set. The set will be used to monitor the risk status of the financial market, provide real-time risk alerts and preventive measures, and the effectiveness of key risk indicators. The calculation formula is as follows:

[0054] in, represents the score of the risk indicator, The statistical significance of the representative indicators ensures that the overall evaluation of the risk indicator set is based on a comprehensive consideration of statistical significance and practical impact.

[0055] Specifically, if Figure 7 As shown in the figure, combining the data in the text risk indicator set with the risk assessment results, analyzing the matching degree between the keywords and the current risk model, optimizing the correlation analysis between the keywords and the risk level, and establishing the automobile financial risk identification model are as follows: S601: According to the text risk indicator set, data is collected, risk indicator data related to automobile finance is integrated, keywords in the text data are classified and relevance is marked, the marked keywords are used to perform key verification on the risk assessment model, and data matching and preliminary screening are performed to generate preliminary keyword mapping data. The execution process is as follows; According to the text risk indicator set, data collection is carried out, which involves automated scripts and API calls to obtain risk data related to auto finance from multiple financial databases and network platforms, integrate risk indicator data related to auto finance, and use data integration technologies, such as data warehouse technology, to unify the format and standard of the collected data for subsequent processing. Keywords in text data are classified and relevance is marked. Text analysis tools are used to identify and mark risk-related keywords and phrases. The marked keywords are used to perform critical verification of the risk assessment model. Simulation tests are used to test the effectiveness and accuracy of keywords in risk prediction. Data matching and preliminary screening are carried out. The marked keywords are matched with known risk indicators using data matching algorithms to generate preliminary keyword mapping data, and keyword tags are matched with risks. The calculation formula is as follows:

[0056] in, Represents keywords, Represents the relevance score of the keyword. represents the risk weight, Represents the number of keywords.

[0057] S602: Quantitatively analyze the relationship between keywords and risk levels using the preliminary keyword mapping data, remap the risk weights of keywords by statistically analyzing the frequency and correlation, including redefining the connection rules between keywords and risk categories, adjusting the output parameters and matching the risk rating requirements, and constructing a keyword risk association mapping table. The execution process is as follows; Utilize the preliminary keyword mapping data to quantitatively analyze the relationship between keywords and risk levels. Determine the correlation between the frequency of occurrence of keywords and their risk levels through statistical methods such as regression analysis. Remap the risk weights of keywords through statistics of frequency and correlation. Use statistical software to perform data analysis and model adjustments to more accurately reflect the risk contribution of each keyword, including redefining the connection rules between keywords and risk categories, adjusting the mapping logic of keywords to risk categories based on the analysis results, adjusting the output parameters and matching the risk rating requirements, adjusting the model output through optimization algorithms to ensure that the output data meets the risk rating standards, building a keyword risk association mapping table, and adjusting the risk weights. The calculation formula is as follows:

[0058] in, Represents the frequency of occurrence of keywords. Represents the risk relevance weight of the keyword, Represents the total number of keywords.

[0059] S603: Based on the keyword risk association mapping table, the interaction between risk indicators and keywords is optimized by adjusting model parameters, including adjusting the model structure and updating the risk weights. The execution process of obtaining the automobile financial risk identification model is as follows; Based on the keyword risk association mapping table, the interaction between risk indicators and keywords is optimized by adjusting model parameters, including using machine learning technology to adjust the model parameter settings to better capture the complex relationship between keywords and risk indicators, including the adjustment of model structure and the update of risk weights. By analyzing the matching degree between the model output and actual risk events, the model structure is optimized to obtain an automobile financial risk identification model. This model will be able to accurately predict and identify potential financial risks, provide strong data support for risk management, and optimize the model. The calculation formula is as follows:

[0060] in, Represents the importance of risk indicators, Represents the matching value between the keyword and the risk indicator. Represents the number of indicators, Optimize the coefficients for the model.

[0061] like Figure 8 As shown, a highway construction optimization system based on big data includes: The data capture module collects data on market interest rates, borrower credit records, and loan terms based on the needs of the auto finance market. Through multi-dimensional correlation analysis, it identifies key influencing factors and performs cluster analysis to obtain a data relationship set between the market and borrowers. The market behavior analysis module uses the data relationship set between the market and borrowers to analyze the correlation between borrower responses and changes in market conditions, adopts change point detection, identifies the shift in behavior patterns, and generates a behavioral market interaction framework; The dynamic interaction model module is based on the behavioral market interaction framework, monitors the changes in market and borrower data in real time, dynamically adjusts the weights in the relationship model, matches market changes, and builds a market dynamic response model; The risk prediction module applies the market dynamic response model to simulate the changes in market and borrower behavior in the future time period, adjusts the risk level according to the model output, adjusts the risk assessment parameters, and obtains the future behavior and risk prediction model; The risk identification and optimization module is based on the future behavior and risk prediction model, analyzes the text data of the loan application, extracts key information points associated with risks, analyzes the information points, optimizes risk ratings, and establishes an auto finance risk identification model.

[0062] The above are only preferred embodiments of the present invention and are not intended to limit the present invention in other forms. Any technician familiar with the profession may use the technical contents disclosed above to change or modify them into equivalent embodiments with equivalent changes and apply them to other fields. However, any simple modification, equivalent change and modification made to the above embodiments based on the technical essence of the present invention without departing from the technical solution of the present invention still falls within the protection scope of the technical solution of the present invention.

Claims

1. An intelligent identification method for automobile financial risks, characterized in that: The following steps are involved: Based on the needs of the auto finance market, we collect data on market interest rates, borrower credit records, and loan terms, analyze the relationships between the data, and form a behavioral market relationship framework; According to the key variables in the behavioral market relationship framework, the changes in market data and borrower credit data are tracked in real time, and the connection weights of the internal variables of the model are dynamically adjusted to create a market dynamic interaction model; Using the market dynamic interaction model, simulate the market conditions and borrower behavior changes in the future time period, adjust the model parameters and match the new market conditions, and build a future behavior prediction adjustment model; By adjusting the future behavior prediction model, the default probability and credit risk of the borrower are regularly updated, the risk level is adjusted, and the risk assessment results are generated; According to the risk assessment results, analyzing the text data in the loan application, extracting keywords and phrases associated with financial risks, refining the extraction process of key information points, and forming a text risk indicator set; Combined with the data in the text risk indicator set and the risk assessment results, the matching degree between the keywords and the current risk model is analyzed, the correlation analysis between the keywords and the risk level is optimized, and an automobile financial risk identification model is established.

2. The intelligent identification method of automobile financial risks according to claim 1 is characterized in that: The behavioral market relationship framework includes interaction mode identification, market behavior response, and periodic changes. The market dynamic interaction model includes weight adjustment indicators, monitoring cycle settings, and data feedback mechanisms. The future behavior prediction adjustment model includes simulation result accuracy, parameter adaptability, and market contingency strategies. The risk assessment results include credit rating updates, default probability corrections, and risk response strategies. The text risk indicator set includes risk semantic analysis, keyword relevance, and information sensitivity. The auto finance risk identification model includes risk level redefinition, keyword utility testing, and model adaptability enhancement.

3. The intelligent identification method of automobile financial risks according to claim 1 is characterized in that: Based on the needs of the auto finance market, the steps to collect market interest rates, borrower credit records and loan terms data, analyze the relationship between the data, and form a behavioral market relationship framework are as follows: Based on the needs of the auto finance market, we collect raw data on market interest rates, borrower credit records, and loan terms, classify and label the data, extract key fields, unify the structure of the data set, adjust the data fields, and establish a unified data set; Based on the unified data set, data records at different time points are marked, data segmentation is performed, time series are associated with borrower behavior data, and the interaction between borrower behavior and market conditions is analyzed through association to construct a time difference data relationship graph; Using the time difference data relationship diagram, the interaction pattern of the data at differentiated time points is analyzed, the data relationships are compared, the behavioral market relationship of the borrower's interaction with market conditions is identified, and the behavioral market relationship framework is obtained.

4. The intelligent identification method of automobile financial risks according to claim 1 is characterized in that: According to the key variables in the behavioral market relationship framework, the changes in market data and borrower credit data are tracked in real time, and the connection weights of the internal variables of the model are dynamically adjusted. The specific steps for creating a market dynamic interaction model are as follows: According to the behavioral market relationship framework, market dynamic data and borrower credit information are collected in real time, including stock market fluctuation data, commodity trading volume and updates on borrowers' current credit ratings, and synchronized to a central database through a network interface to obtain market credit data; Based on the market credit data, dynamically adjust the connection weights between the internal variables of the model, refer to the frequency and amplitude of data changes and current data input, update the weights in real time through the internal adjustment mechanism, and generate a weight adjustment model; Based on the weight adjustment model, a machine learning algorithm is used to simulate and predict the market dynamics of the interaction between market behavior and borrower credit status, collect and organize market behavior data, evaluate borrower credit status, perform model verification and adjustment, and obtain a market dynamic interaction model.

5. The intelligent identification method of automobile financial risks according to claim 4 is characterized in that: The formula for the machine learning algorithm is as follows: in, represents the predicted output of the model, is the activation function, is the weight matrix, is the input feature vector, is the bias term, is the adjustment factor, The degree of market competition, is the risk adjustment factor, Score industry risk.

6. The intelligent identification method of automobile financial risk according to claim 1 is characterized in that: Using the market dynamic interaction model, simulating the market conditions and borrower behavior changes in the future time period, adjusting the model parameters and matching the new market conditions, the steps of constructing the future behavior prediction adjustment model are as follows: Using the market dynamic interaction model, simulate changes in market conditions and borrower behavior in future time periods, record key indicators in the output, including interest rate fluctuations and changes in borrowing intentions, time-mark the indicators, classify trends in differentiated market behaviors, and generate market forecast data records; Based on the market forecast data records, identifying the change points in the parameters, adjusting the influencing parameters in the model, including adjusting the credit score weights, interest rate response sensitivity, recalibrating for new market data, matching the current changes in the market, and obtaining an optimized parameter configuration; According to the optimization parameter configuration, the logical structure of the behavior prediction model in the future time period is updated, the prediction logic is adjusted and the model's matching ability to market changes in the future time period is optimized, the interaction mode between parameters is refined, and a future behavior prediction adjustment model is constructed.

7. The intelligent identification method of automobile financial risk according to claim 1 is characterized in that: The steps of regularly updating the borrower's default probability and credit risk, adjusting the risk level, and generating the risk assessment results by using the future behavior prediction adjustment model are as follows: Extracting the default probability and credit risk data of each borrower from the future behavior prediction adjustment model, including account behavior analysis and payment record update, and automatically synchronizing the data from the market monitoring configuration to the risk assessment database on a daily basis to obtain a default probability data set; Based on the default probability data set, analyze the current market data changes, adjust the risk level of the borrower item by item, compare the current market dynamics with historical credit behavior, automatically adjust the risk level threshold, and generate an updated risk level; The updated risk level is used in combination with current market data to conduct risk assessment, and the data of each borrower is analyzed and evaluated through credit assessment to obtain a risk assessment result.

8. The intelligent identification method of automobile financial risks according to claim 1 is characterized in that: According to the risk assessment results, the text data in the loan application is analyzed, keywords and phrases associated with financial risks are extracted, and the extraction process of key information points is refined to form a text risk indicator set. Specifically, the steps are as follows: Based on the risk assessment results, words and phrases related to financial risks are screened from the loan application text, words are identified by text tagging technology, financial terms, risk descriptions and related indicators are classified and sorted, and a risk vocabulary library is constructed; According to the risk vocabulary library, adjust the word frequency analysis parameters, strengthen the context association analysis, optimize the word risk weight, refine the risk indicator extraction, and generate a risk information set; Through the risk information set, key risk indicators are integrated, related risk data are collected and analyzed, key risk indicators are screened and determined, and a text risk indicator set is generated.

9. The intelligent identification method of automobile financial risks according to claim 1 is characterized in that: Combining the data in the text risk indicator set with the risk assessment results, analyzing the matching degree between the keywords and the current risk model, optimizing the correlation analysis between the keywords and the risk level, and establishing the automobile financial risk identification model in the following specific steps: According to the text risk indicator set, data is collected, risk indicator data related to automobile finance is integrated, keywords in the text data are classified and relevance is marked, the marked keywords are used to perform key verification on the risk assessment model, and data matching and preliminary screening are performed to generate preliminary keyword mapping data; Utilizing the preliminary keyword mapping data, quantitatively analyzing the relationship between keywords and risk levels, remapping the risk weights of keywords through statistics of frequency and relevance, including redefining the connection rules between keywords and risk categories, adjusting output parameters and matching risk rating requirements, and constructing a keyword risk association mapping table; Based on the keyword risk association mapping table, the interaction between risk indicators and keywords is optimized by adjusting model parameters, including adjusting the model structure and updating the risk weights, to obtain an automobile financial risk identification model.

10. An intelligent identification system for automobile financial risks, characterized in that: According to the method for intelligently identifying automobile financial risks according to any one of claims 1 to 9, the system comprises: The data capture module collects market interest rates, borrower credit records, and loan terms data based on the needs of the auto finance market, identifies key influencing factors through association analysis, and obtains a data relationship set between the market and borrowers; The market behavior analysis module uses the data relationship set between the market and the borrower to analyze the correlation between the borrower and the market conditions, adopts change point detection, and generates a behavioral market interaction framework; The dynamic interaction model module is based on the behavioral market interaction framework, monitors the changes in market and borrower data in real time, dynamically adjusts the weights in the relationship model, matches market changes, and builds a market dynamic response model; The risk prediction module applies the market dynamic response model to simulate the changes in market and borrower behavior in the future time period, adjusts the risk level according to the model output, adjusts the risk assessment parameters, and obtains the future behavior and risk prediction model; The risk identification optimization module analyzes the text data of the loan application based on the future behavior and risk prediction model, extracts key information points associated with risks, analyzes the information points, optimizes risk ratings, and establishes an automobile financial risk identification model.

Citation Information

Cited By

  • Automobile financial risk assessment method and system based on artificial intelligence

    CN119648392A

  • Automobile financial transaction data transmission method, system and equipment based on artificial intelligence

    CN119728667A

  • Method and system for generating financial information graph of user through financing platform

    CN120045883A

  • A method and system for generating a user's financial information graph through a financing platform

    CN120045883B

  • Financial multi-source data processing system and method

    CN120086577A