Disclosure of Invention
The invention aims to solve the defects in the prior art, and provides a post-finance risk early warning treatment method for automobiles.
In order to achieve the purpose, the invention adopts the following technical scheme that the method for disposing the risk early warning after the automobile finance credit comprises the following steps:
s1, acquiring and analyzing bank statement and consumption record information of a borrower based on post-loan transaction information of the borrower, wherein the information comprises the amount, time and frequency of a plurality of transactions, and generating borrower bill data;
S2, based on the borrower bill data, identifying abnormal transaction information and risk transaction behaviors by analyzing transaction objects, time and amount of multiple transactions, and generating a transaction behavior analysis result;
S3, based on the transaction behavior analysis result, evaluating the automobile value by analyzing market trend, vehicle condition and historical use data of the borrower vehicle, and generating a mortgage asset evaluation result by combining credit record and economic condition of the borrower;
s4, calculating repayment risk scores by analyzing asset conditions of a plurality of borrowers and combining historical repayment behaviors and overdue records by using the mortgage asset evaluation results, and generating risk borrower identification results;
S5, matching the notification information of the charge-back for the borrowers with a plurality of risk levels according to the identification result of the risk borrowers, adjusting the sending frequency of the notification, recording the repayment response behaviors of the borrowers, and generating a charge-back response record;
And S6, using the collect response record, considering the financial condition and the fund recovery efficiency of the borrower, adjusting the repayment period and the overdue interest rate of the loan, calculating the repayment amount, and generating a post-loan risk processing result.
As a further aspect of the present invention, the borrower bill data includes transaction date information, transaction object information, and transaction amount information, the transaction behavior analysis result includes abnormal transaction identification records, risk transaction type data, and risk transaction period information, the mortgage asset evaluation result includes real-time automobile market value evaluation data, expected depreciation rate, maintenance, and repair records, the risk borrower identification result includes credit risk rating, predicted default rate, and borrower financial pressure evaluation result, the collect response record includes response time analysis results, repayment amount change information, and repayment willingness evaluation information, and the post-credit risk processing result includes repayment plan adjustment results, overdue interest rate update results, and expected recovery efficiency of funds.
As a further aspect of the present invention, collecting and analyzing bank running bills and consumption record information of a borrower based on post-loan transaction information of the borrower, including amounts, time and frequency of a plurality of transactions, the step of generating borrower bill data specifically includes:
S101, collecting bank running bills and consumption records of borrowers based on post-loan transaction information of the borrowers, recording transaction date, amount and frequency information, and generating a transaction data collection result;
s102, cleaning and formatting data based on the transaction data acquisition result, identifying repeated records and error input, and generating a consumption record processing result;
And S103, classifying transaction data based on the consumption record processing result, calculating total amount and transaction frequency of various transactions, and generating borrower bill data.
As a further scheme of the present invention, based on the borrower bill data, by analyzing the transaction object, time and amount of multiple transactions, identifying abnormal transaction information and risk transaction behaviors, the step of generating a transaction behavior analysis result specifically includes:
S201, based on the borrower bill data, identifying abnormal transaction amount and transaction time by analyzing the amount and transaction time of multiple transactions and comparing the amount and the transaction time with a preset monitoring threshold value, and generating an abnormal transaction comparison result;
s202, analyzing background information of abnormal transactions, including payment sites and transaction objects, classifying and recording risk transactions based on the abnormal transaction comparison result, and generating a risk transaction classification result;
s203, according to the risk transaction classification result, calculating credit influence scores of various transaction behaviors by evaluating influences of the various transaction behaviors on credit states of borrowers and adopting a logistic regression algorithm, and generating a transaction behavior analysis result.
As a further aspect of the present invention, the logistic regression algorithm is according to the formula:
Calculating the default probability of the borrower, wherein p is the default probability of the borrower, b0 is an intercept term, b1 is a weight related to transaction amount, b2 is a weight related to transaction frequency, b3 is a weight related to transaction time interval, b4 is a weight related to historical overdue times of the borrower, b5 is a weight related to funds inflow in the borrower account, x1 is transaction amount, x2 is transaction frequency, x3 is transaction time interval, x4 is historical overdue times of the borrower, x5 is funds inflow in the borrower account, and e is a natural logarithm base.
As a further scheme of the present invention, based on the transaction behavior analysis result, the market trend, the vehicle condition and the historical usage data of the borrower vehicle are analyzed to evaluate the automobile value, and the step of generating the mortgage asset evaluation result by combining the credit record and the economic condition of the borrower is specifically as follows:
S301, collecting vehicle information of borrowers based on the transaction behavior analysis result, recording basic conditions and historical use data of vehicles, including vehicle types, vehicle ages, mileage and maintenance records, and generating vehicle condition data records;
s302, analyzing sales prices and demand trends of various automobiles through real-time market data based on the automobile condition data record, and generating a market trend analysis result;
S303, according to the market trend analysis result, the vehicle value and the change trend of the borrower are evaluated in real time, and the mortgage asset evaluation result is generated by combining the credit record and the economic condition of the borrower.
As a further scheme of the invention, by using the mortgage asset evaluation result and analyzing the asset conditions of a plurality of borrowers and combining the historical repayment behaviors and overdue records, the repayment risk score is calculated, and the step of generating the risk borrower identification result specifically comprises the following steps:
s401, acquiring and recording asset condition data and historical repayment behavior records of a plurality of borrowers by using the mortgage asset evaluation result, wherein the historical repayment behavior records comprise overdue records and repayment frequency information, and generating asset and behavior data;
s402, analyzing historical repayment behaviors and overdue records of borrowers based on the asset and behavior data, evaluating stability and overdue probability of the repayment behaviors, and generating a stability analysis result;
s403, calculating repayment risk scores of a plurality of borrowers based on the stability analysis result, and evaluating overdue risk grades in consideration of asset values and credit conditions to generate a risk borrower identification result.
As a further scheme of the present invention, according to the identification result of the risk borrower, the present invention matches the notification information of the induction and collection for the borrowers with multiple risk levels, adjusts the sending frequency of the notification, records the repayment response behavior of the borrower, and generates the induction and collection response record specifically as follows:
S501, matching the notification content of the notification of the collection according to the risk borrower identification result and the risk grades of a plurality of borrowers, wherein the notification content comprises notification information and sending frequency, and generating a matching result of the notification of the collection;
s502, based on the result of the prompt notification matching, sending prompt notification information to a plurality of borrowers, recording repayment behaviors of the borrowers after receiving the notification, including repayment time and amount, and generating response data of the prompt notification information;
S503, analyzing the response data of the information to be collected, and adopting a K-means clustering algorithm to evaluate the effects of various notification contents of collection by analyzing the response speed and the repayment completion degree of borrowers so as to generate a collection response record.
As a further scheme of the invention, the K-means clustering algorithm is as follows:
Calculating a clustering effect of the furcation strategy, wherein S is total internal variance, k is the number of clusters, Ci is a data point set in the ith cluster, xtime is a response time data point of the borrower, xamount is a response amount data point of the borrower, xbehavior is a repayment behavior data point of the borrower, mutime is a center point of a time cluster, muamount is a center point of an amount cluster, mubehavior is a center point of a behavior cluster, w1 is a time weight coefficient, w2 is an amount weight coefficient, w3 is a behavior weight coefficient, i is a cluster index, and x is a single data point.
As a further scheme of the invention, the step of using the collect response record to consider the financial condition and the fund recovery efficiency of the borrower, adjusting the repayment period and the overdue interest rate of the loan and calculating the repayment amount, and generating the post-loan risk processing result specifically comprises the following steps:
S601, analyzing the collect response records, and identifying overdue risks of a plurality of borrowers by analyzing the repayment amounts and loan interest rates of the plurality of borrowing contracts and comparing the pay amounts and the loan interest rates with actual repayment capacities of the plurality of borrowers to generate overdue risk identification data;
s602, identifying a repayment period and a overdue interest rate to be adjusted based on the overdue risk identification data in consideration of the fund recovery efficiency and the financial capability of the borrower, and generating repayment condition adjustment data;
s603, adjusting the repayment contract of the borrower according to the repayment condition adjustment data, monitoring the adjusted repayment behavior and analyzing the influence on repayment efficiency, and generating a post-loan risk processing result.
Compared with the prior art, the invention has the advantages and positive effects that:
According to the invention, through the consumption records of borrowers and bank running bills, transaction objects, time and amounts of a plurality of transactions are analyzed, risk transaction behaviors are identified, default risks are identified in early stage, the value of mortgage assets is evaluated by combining the market trend of vehicles and the credit records of the borrowers, the accuracy of asset evaluation is improved, the overdue repayment period is adjusted according to the actual repayment capability of the borrowers by calculating repayment risk scores and matching with the repayment notification, the fund recovery rate and the financial stability of loan service are improved, the risk management efficiency is optimized, the loss caused by default is reduced, and the benefits of consumers and investors are protected.
Detailed Description
The present invention will be described in further detail with reference to the drawings and examples, in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
In the description of the present invention, it should be understood that the terms "length," "width," "upper," "lower," "front," "rear," "left," "right," "vertical," "horizontal," "top," "bottom," "inner," "outer," and the like indicate orientations or positional relationships based on the orientation or positional relationships shown in the drawings, merely to facilitate describing the present invention and simplify the description, and do not indicate or imply that the devices or elements referred to must have a specific orientation, be configured and operated in a specific orientation, and therefore should not be construed as limiting the present invention. Furthermore, in the description of the present invention, the meaning of "a plurality" is two or more, unless explicitly defined otherwise.
Referring to fig. 1, the present invention provides a technical solution, a post-financial risk early warning treatment method for an automobile, comprising the following steps:
s1, acquiring and analyzing bank statement and consumption record information of a borrower based on post-loan transaction information of the borrower, wherein the information comprises the amount, time and frequency of a plurality of transactions, and generating borrower bill data;
S2, based on borrower bill data, identifying abnormal transaction information and risk transaction behaviors by analyzing transaction objects, time and amount of multiple transactions, and generating a transaction behavior analysis result;
S3, based on the transaction behavior analysis result, evaluating the automobile value by analyzing market trend, automobile condition and historical use data of the borrower automobile, and generating a mortgage asset evaluation result by combining the credit record and economic condition of the borrower;
s4, calculating repayment risk scores by analyzing asset conditions of a plurality of borrowers and combining historical repayment behaviors and overdue records by using mortgage asset assessment results, and generating risk borrower identification results;
s5, matching the notification information of the charge-back for the borrowers with a plurality of risk levels according to the identification result of the risk borrowers, adjusting the sending frequency of the notification, recording the repayment response behaviors of the borrowers, and generating a charge-back response record;
S6, using the collect response record, considering the financial condition and the fund recovery efficiency of the borrower, adjusting the repayment period and the overdue interest rate of the loan, calculating the repayment amount, and generating a post-loan risk processing result.
The borrower bill data comprises transaction date information, transaction object information and transaction amount information, the transaction behavior analysis result comprises abnormal transaction identification records, risk transaction type data and risk transaction period information, the mortgage asset assessment result comprises real-time automobile market value assessment data, expected depreciation rate, maintenance and repair records, the risk borrower identification result comprises credit risk rating, predicted default rate and borrower financial pressure assessment result, the furcation response record comprises response time analysis results, repayment amount change information and repayment willingness assessment information, and the post-credit risk processing result comprises repayment plan adjustment results, overdue interest rate update results and expected fund recovery efficiency.
Referring to fig. 2, based on the post-credit transaction information of the borrower, the steps of collecting and analyzing bank running bills and consumption record information of the borrower, including the amount, time and frequency of a plurality of transactions, and generating the borrower bill data are specifically as follows:
S101, collecting bank running bills and consumption records of borrowers based on the post-loan transaction information of the borrowers, and recording transaction date, amount and frequency information, wherein the process for generating a transaction data collection result is specifically as follows;
In the S101 substep, based on the post-loan transaction information of the borrower, collecting bank running bills and consumption records of the borrower, including exporting running data from a bank system through an API (application program interface), integrating the consumption records of an e-commerce platform, capturing transaction date, amount and frequency information through a data crawler technology, verifying the validity and integrity of the data, and ensuring that the time stamp and amount of each transaction record are verified, wherein the formula is as follows:
Wherein Tioz represents the adjusted transaction time coefficient, Aioz represents the original transaction amount, kz is a time coefficient adjustment parameter, and is used for normalizing the time value in the transaction record, betaz is an offset of time coefficient adjustment, and is used for normalizing the time and amount of the transaction record, ensuring the consistency and comparability of data, and generating a transaction data acquisition result.
S102, cleaning and formatting data based on a transaction data acquisition result, and identifying repeated records and error input, wherein the process of generating a consumption record processing result is specifically as follows;
In the step S102, based on the transaction data collection result, the data is cleaned and formatted, including using the custom data cleaning script to identify and reject repeated items and error inputs in the transaction record, locating the repeated transaction record by hash table technology, automatically correcting obvious input errors by predefined rules, and standardizing the transaction data into a unified format, where the formula is:
Wherein Djz represents the quality index after data cleaning, Xkoz represents the amount of a single transaction,Is the average value of all transaction amounts, and n0 is the total number of transactions, is used for evaluating the effect of data cleaning, ensuring the consistency and the accuracy of a data set and generating a consumption record processing result.
S103, classifying transaction data based on a consumption record processing result, and calculating total amount and transaction frequency of various transactions, wherein the process of generating borrower bill data is specifically as follows;
In the step S103, based on the processing result of the consumption record, classifying the transaction data, including using a decision tree algorithm to distinguish the transactions of different types, calculating the total amount and the transaction frequency of each type of transaction, understanding the pattern of the consumption behavior and predicting the consumption trend, wherein the formula is:
Wherein, Ckz represents the average transaction cost of a certain category, floz and aloz are the frequency and the amount of the first0 transaction, m0 is the total amount of the target category transaction, and the consumption intensities of different categories are calculated and compared to generate borrower bill data.
Referring to fig. 3, based on borrower bill data, by analyzing transaction objects, time and amount of multiple transactions, abnormal transaction information and risk transaction behaviors are identified, and the steps of generating a transaction behavior analysis result are specifically as follows:
S201, based on borrower bill data, identifying abnormal transaction amount and transaction time by analyzing the amount and transaction time of multiple transactions and comparing the amount and the transaction time with a preset monitoring threshold, wherein the process for generating an abnormal transaction comparison result is specifically as follows;
In the step S201, based on the borrower bill data, the amount and the transaction time of the multiple transactions are analyzed and compared with the preset monitoring threshold, and the following formula is used:
Wherein Ei1x represents a transaction anomaly score, Ai1x represents the amount of a single transaction, θAx is a preset monitoring threshold for the amount, σAx is the standard deviation of the amount, Ti1x represents transaction time, θTx is a preset monitoring threshold for time, σTx is the standard deviation of time, and transactions deviating from the normal mode are identified to generate an anomaly transaction comparison result.
S202, analyzing background information of abnormal transactions based on abnormal transaction comparison results, wherein the background information comprises payment sites and transaction objects, classifying and recording risk transactions, and the process of generating risk transaction classification results is specifically as follows;
in the step S202, based on the comparison result of the abnormal transaction, the background information of the abnormal transaction is analyzed, including the payment location and the transaction object, and the occurrence frequency and the influence of each type of risk transaction are recorded by using the classification result, wherein the formula is as follows:
Wherein Rk1x represents risk scores of target classifications, wpj1x and Woj1x are risk weight coefficients of payment locations and transaction objects respectively, ppj1x and Poj1x are risk probability scores calculated according to the payment locations and the transaction objects, n1 is total number of transactions in the classifications, the risk scores of the payment locations and the transaction objects of each transaction are summed up by means of weighted summation, classification results of risk transactions are evaluated and recorded, and risk transaction classification results are generated.
S203, calculating credit influence scores of various transaction behaviors by evaluating the influence of the various transaction behaviors on the credit state of the borrower according to the risk transaction classification result and adopting a logistic regression algorithm, wherein the process for generating the transaction behavior analysis result is specifically as follows;
In step S203, according to the risk transaction classification result, the influence of multiple transaction behaviors on the credit state of the borrower is evaluated, a logistic regression algorithm is adopted, and the transaction amount, the transaction frequency, the time interval of the transaction, the historical overdue times of the borrower and the funds inflow information in the borrower account of multiple transactions are combined to calculate the credit influence score of each behavior, so that the influence of different transaction behaviors on the credit state of the borrower is quantified, and a transaction behavior analysis result is generated.
Logistic regression algorithm, according to the formula:
Calculating the default probability of the borrower, wherein p is the default probability of the borrower, b0 is an intercept term, b1 is a weight related to transaction amount, b2 is a weight related to transaction frequency, b3 is a weight related to transaction time interval, b4 is a weight related to historical overdue times of the borrower, b5 is a weight related to funds inflow in the borrower account, x1 is transaction amount, x2 is transaction frequency, x3 is transaction time interval, x4 is historical overdue times of the borrower, x5 is funds inflow in the borrower account, and e is a natural logarithm base.
Calculating the default probability of the borrower, wherein p is the probability of the borrower default, b0 is the intercept term, the default value under the influence of no input variable when the model output is provided, b1 is the weight related to the transaction amount, the influence of the amount on the default probability is measured, b2 is the weight related to the transaction frequency, the influence of the transaction frequency on the default probability is measured, b3 is the weight related to the transaction time interval, the influence of the time interval on the default probability is measured, b4 is the weight related to the historical overdue number of the borrower, the influence of the historical overdue action on the default probability is measured, b5 is the weight related to the funds influx in the borrower account, the influence of the default probability is measured, x1 is the transaction amount, the scale representing the transaction, x2 is the transaction frequency, the number of transactions actively used by the borrower, x3 is the time interval describing the transaction, the time of occurrence, x is the credit withdrawal amount used for the borrower, and the historical overdue number of the borrower account is the financial overdue number used for reflecting the default number of the borrower account, and the historical overdue number of the borrower account is used for the financial overdue number.
The specific implementation process of the formula is as follows:
Parameters of transaction amount, transaction frequency and time interval of transaction are obtained through bank flow data of the borrower, parameters of historical overdue times of the borrower are obtained through credit reports, parameters of funds inflow in an account of the borrower are obtained through account flow of the borrower, b0、b1、b2、b3、b4、b5 parameters are obtained through a logistic regression model, the parameters are substituted into a formula, the probability of default of the borrower is calculated, and the method is used for evaluating and managing default risks of the borrower.
Referring to fig. 4, based on the transaction behavior analysis result, the market trend, the vehicle condition and the historical usage data of the borrower vehicle are analyzed to evaluate the automobile value, and the steps of generating the mortgage asset evaluation result by combining the credit record and the economic condition of the borrower are specifically as follows:
s301, collecting vehicle information of borrowers based on transaction behavior analysis results, and recording basic conditions and historical use data of vehicles, wherein the basic conditions and the historical use data comprise vehicle types, vehicle ages, mileage and maintenance records, and the flow of generating vehicle condition data records is specifically as follows;
In the step S301, based on the analysis result of the transaction behavior, the basic condition and the historical usage data of the vehicles are collected through the data, the model, the age, the mileage and the maintenance record of each vehicle are recorded, the format standardization and the data integrity check are performed on the data, the data accuracy is ensured, and the formula is:
Wherein Vi2c represents the vehicle comprehensive condition score, alphac is the weight coefficient of the vehicle age, betac is the weight coefficient of mileage, gammac is the weight coefficient of maintenance times, Ai2c is the vehicle age, Di2c is mileage, Rk2c is a single maintenance record, m2 is the number of maintenance records, and the overall condition of the vehicle is quantified by integrating the influences of the vehicle age, mileage and maintenance records to generate a vehicle condition data record.
S302, analyzing sales prices and demand trends of various automobiles through real-time market data based on the vehicle condition data record, wherein the flow for generating a market trend analysis result is specifically as follows;
In the step S302, based on the vehicle condition data record, the vehicle condition data record is connected to an automobile sales platform and an online market data source in real time, and sales prices and demand trends of different vehicle types are acquired and analyzed by utilizing data mining, pattern recognition and data grabbing algorithms, wherein the formulas are as follows:
Wherein Sj2c is a market trend score, lambdac is a weight coefficient of price, muc is a weight coefficient of demand trend, Pj2c is sales price, Tj2c is demand trend, and the performances and potential values of different vehicle types in the market are evaluated to generate a market trend analysis result.
S303, according to market trend analysis results, real-time evaluating the vehicle value and variation trend of the borrower, and combining the credit record and economic condition of the borrower, wherein the process for generating mortgage asset evaluation results is specifically as follows;
In the step S303, according to the market trend analysis result, the vehicle value and the variation trend of the borrower are evaluated in real time, and by combining the vehicle condition data and the market trend data with the credit record and the economic condition of the borrower, the influence of the values of various vehicles is analyzed by using a financial modeling and credit scoring algorithm, wherein the formula is as follows:
Wherein Zk2c is a mortgage asset assessment result, vc is a weight coefficient of a vehicle condition score, ζc is a weight coefficient of a market trend score, ωc is a weight coefficient of a credit record, ψc is a weight coefficient of an economic condition, Vk2c is a vehicle condition score, Sk2c is a market trend score, Ck2c is a credit record, Ek2c is an economic condition, and the mortgage value of the vehicle is assessed by combining the vehicle condition, market performance, borrower credit and economic condition to generate a mortgage asset assessment result.
Referring to fig. 5, using mortgage asset assessment results, by analyzing asset conditions of a plurality of borrowers and combining historical repayment behaviors and overdue records, a repayment risk score is calculated, and the step of generating a risk borrower identification result specifically includes:
S401, acquiring and recording asset condition data and historical repayment behavior records of a plurality of borrowers by using mortgage asset assessment results, wherein the processes for generating asset and behavior data comprise overdue records and repayment frequency information;
In the step S401, using the mortgage asset assessment result, collecting and recording asset status data and historical repayment behavior records of a plurality of borrowers, extracting asset information of the borrowers from a financial institution database, including the value of the property and the vehicle, combining the historical repayment behavior data, including the overdue records and repayment frequency, performing timestamp marking and validity verification on the data, and ensuring the accuracy and integrity of the records, wherein the formula is as follows:
Wherein Ai3v represents the adjusted asset condition index, Vi3v represents the original asset value, and cv and pv are adjustment parameters for normalizing the asset value to generate asset and behavior data.
S402, analyzing historical repayment behaviors and overdue records of borrowers based on asset and behavior data, and evaluating stability and overdue probability of the repayment behaviors, wherein the process for generating a stability analysis result is specifically as follows;
in the step S402, based on the asset and behavior data, the historical repayment behaviors and overdue records of the borrowers are analyzed, the stability and overdue probability of the repayment behaviors of each borrower are evaluated through regression analysis, a prediction model is built through analysis of the historical data, and the default risk is identified according to the formula:
Wherein Sj3v represents a stability index of the repayment behavior, Rk3v represents a repayment record of the kth3 borrower,The average value of repayment records of all borrowers is n3, the total number of borrowers is n3, the stability of repayment behaviors is calculated, and a stability analysis result is generated.
S403, calculating repayment risk scores of a plurality of borrowers based on the stability analysis result, and evaluating overdue risk grades in consideration of asset values and credit conditions, wherein the process for generating the risk borrower identification result is specifically as follows;
in the step S403, based on the stability analysis result, calculating repayment risk scores of the plurality of borrowers, and considering the asset value and the credit status of each borrower, evaluating the overdue risk level of each borrower by using a weighted score method, where the formula is as follows:
Wherein, Rl3v represents the repayment risk score, fm3v and wm3v are the risk factor score and the corresponding weight of the m3 borrower, q3 is the total risk factor number estimated, the risk grade of the borrower is estimated and calculated, and the risk borrower identification result is generated.
Referring to fig. 6, according to the identification result of the risk borrower, the steps of matching the notification information for the borrowers with multiple risk levels, adjusting the sending frequency of the notification, recording the repayment response behavior of the borrowers, and generating the notification response record are specifically as follows:
S501, matching the notification content of the notification of the collection according to the risk grade of a plurality of borrowers according to the identification result of the risk borrowers, wherein the notification information and the sending frequency are included, and the flow for generating the matching result of the notification of the collection is specifically as follows;
In S501 substep, according to the identification result of the risk borrower, the notification content of the collection is matched according to the risk levels of multiple borrowers, including notification information and sending frequency, where the formula is:
Wherein, Ci4b represents the effect index of the targeted prompt and receipt notification, Ri4b is the risk level of borrowers, αb、γb and μb are tuning parameters that optimize pertinence and efficiency of the notification of the annunciation, and generating a prompting notification matching result.
S502, based on the result of the notification match, sending notification information to a plurality of borrowers, and recording repayment behaviors of the borrowers after receiving the notification, wherein the repayment behaviors comprise repayment time and amount, and the flow for generating response data of the notification information is specifically as follows;
In the sub-step S502, based on the notification of the match result by the reception-promoting, sending the notification information to a plurality of borrowers, recording the repayment behaviors of the borrowers after receiving the notification, the payment time and the amount are monitored through real-time data, the timeliness and the tracking effect of each notification are optimized, and the formula is as follows:
Wherein, Rj4b represents the average repayment time weighted repayment amount, tk4b and pk4b respectively represent the repayment time and amount of the k4 borrower, n4 is the total number of borrowers receiving the notification, the actual effect of the notification of the prompt is analyzed, and the response data of the prompt information is generated.
S503, analyzing response data of the collection information, and adopting a K-means clustering algorithm to evaluate the effects of various collection notification contents by analyzing the response speed and repayment completion degree of borrowers, wherein the process of generating collection response records is specifically as follows;
In the step S503, the response data of the information of the collection is analyzed, the response time, the response amount and the repayment behavior parameters are obtained by collecting the response data of the borrower through a K-means clustering algorithm, the effect of the notification content of the collection is evaluated, the influence of the notification content of the collection on the behaviors of different borrowers is quantified, and the response record of the collection is generated.
K-means clustering algorithm, according to the formula:
Calculating a clustering effect of a furcation strategy, wherein S is a total internal variance and used for measuring the overall compactness of clusters, k is the number of clusters and used for designating the quantity of clusters, Ci is a data point set in an ith cluster and used for representing the average performance of repayment in each class, xtime is a response time data point of a borrower and used for measuring the timeliness of furcation notification, xamount is a response amount data point of the borrower and used for measuring the economic influence of furcation notification, xbehavior is a repayment action data point of the borrower and used for measuring the behavior effect of furcation notification, mutime is a central point of a time cluster and used for representing the average value of response time in each cluster, muamount is a central point of an amount cluster and used for representing the average value of response amount in each cluster, mubehavior is a central point of behavior cluster and used for representing the average performance of repayment in each cluster, w1 is a time weight coefficient, w2 is an importance of amount in cluster analysis, w3 is a behavior weight coefficient, and used for adjusting the importance of amount in cluster analysis, w3 is a behavior weight coefficient, and used for evaluating the importance of a cluster in a cluster and a cluster is used for evaluating the cluster in a cluster and a cluster is different from the cluster center in a cluster and a cluster in a cluster to be used for evaluating the cluster.
The specific implementation process of the formula is as follows:
The method comprises the steps of collecting the response data of the borrower, obtaining parameters of response time xtime, response amount xamount and repayment behavior xbehavior, determining a weight coefficient w1、w2、w3 by using a statistical method according to historical data and a preset analysis target, adjusting the coefficient through gradient descent to ensure optimization of a clustering result, substituting the parameters into a formula, executing clustering operation by using a K-means clustering algorithm, calculating internal variances S of each cluster, and evaluating influences of different notification of the repayment on the behavior of the borrower, so that financial institutions are helped to optimize the repayment strategy and improve the fund recycling efficiency.
Referring to fig. 7, the steps of using the collect response record to consider the financial status and the funds recovery efficiency of the borrower, adjusting the repayment period and the overdue interest rate of the loan, and calculating the repayment amount, and generating the post-loan risk processing result are specifically as follows:
s601, analyzing the collection response records, and identifying overdue risks of a plurality of borrowers by analyzing the repayment amounts and loan interest rates of the plurality of borrowing contracts and comparing the repayment amounts and the loan interest rates with actual repayment capacities of the plurality of borrowers, wherein the process for generating overdue risk identification data is specifically as follows;
In the step S601, the collect response record is analyzed, and the financial pressure index of each borrower is calculated by analyzing the payoff amount and loan interest rate of the plurality of borrowing contracts and comparing with the actual payoff capability of the plurality of borrowers, and the overdue risk is identified by the formula:
wherein, Ri5q represents a overdue risk index for measuring the possibility of default of the borrower, Li5q represents the loan interest rate of the borrower, reflecting the borrowing cost, Ci5q represents the month income of the borrower, and is used as a direct index of repayment capability, lambdaq is a overdue risk adjustment coefficient for adjusting the influence of the loan interest rate on overdue risk assessment, epsilonq is a constant for preventing denominator from being zero, and overdue risk identification data is generated.
S602, identifying a repayment period and a overdue interest rate to be adjusted based on overdue risk identification data in consideration of fund recovery efficiency and financial capability of borrowers, wherein the process of generating repayment condition adjustment data is specifically as follows;
In the step S602, based on the overdue risk identification data, the funds recovery efficiency and the financial ability of the borrower are considered, the repayment period and the overdue interest rate to be adjusted are identified, and the funds recovery is optimized by calculating the appropriate repayment period and the increased overdue interest rate of each borrower, with the formula:
Wherein, Pj5q represents a recommended repayment cycle for adjusting a repayment plan of a borrower, Fj5q represents a fund recovery efficiency, and a past repayment performance of the borrower is measured, Gj5q represents a financial pressure index of the borrower, omegaq is a cycle adjustment coefficient for adjusting a length of the repayment cycle according to the fund recovery efficiency based on a income and a debt burden calculation of the borrower, and thetaq is a power index of the financial pressure adjustment index for adjusting an influence degree of the financial pressure on the repayment cycle to generate repayment condition adjustment data.
S603, adjusting the repayment contract of the borrower according to the repayment condition adjustment data, monitoring the adjusted repayment behavior and analyzing the influence on repayment efficiency, wherein the flow of generating the post-loan risk processing result is specifically as follows;
In the step S603, the payment contract of the borrower is adjusted according to the payment condition adjustment data, including modifying the payment period and overdue interest rate in the contract, monitoring the adjusted payment behavior and analyzing the influence on the payment efficiency, where the formula is:
Wherein Ek5q represents an improvement index of repayment efficiency for evaluating the repayment behavior improvement degree after contract adjustment, Hk5q represents adjusted repayment behavior data including the frequency and amount of on-time repayment, sigmaq is an efficiency improvement coefficient for enlarging or reducing the evaluation result of repayment behavior improvement, ρq is an adjustment constant for data stability, ensuring the feasibility of logarithmic calculation, evaluating the influence of the adjusted repayment condition on the overall financial health, and generating a post-loan risk processing result.
The present invention is not limited to the above embodiments, and any equivalent embodiments which can be changed or modified by the technical disclosure described above can be applied to other fields, but any simple modification, equivalent changes and modification made to the above embodiments according to the technical matter of the present invention will still fall within the scope of the technical disclosure.