An internet financial risk assessment system and method based on multi-dimensional data fusion

Abstract

The application relates to the field of financial risk assessment, in particular to an internet financial risk assessment system and method based on multi-dimensional data fusion. The method comprises the following steps: setting an internet data point, acquiring user financial data, user information data and a user timestamp, and constructing a user financial tree; monitoring the user financial tree, acquiring a monitoring timestamp and financial monitoring data, obtaining a user radius, a financial orientation value and a multi-dimensional financial fusion graph according to the monitoring timestamp and the financial monitoring data; analyzing the user financial tree and the multi-dimensional financial fusion graph through the user timestamp, obtaining a change time value, obtaining a risk trend value and a trend time value according to the multi-dimensional financial fusion graph and the change time value; and predicting through the risk trend value and the trend time value, obtaining a predicted risk trend value and a predicted trend time value, and assessing the financial risk of the user. The application can improve the user's financial risk response ability.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of financial risk assessment, in particular to an internet financial risk assessment system and method based on multi-dimensional data fusion. BACKGROUND

[0002] With the increasing complexity of global economy and financial markets, the types of risks faced by enterprises are increasingly diversified. In the field of internet consumer finance, with the rapid growth of online lending demand, financial risk management becomes particularly important. These risks not only include traditional credit risk, market risk and liquidity risk, but also involve the influence of external factors such as macroeconomic, industry changes and policy changes. The data sources are increasingly diversified, and the types and forms of data are very rich.

[0003] The method, system and device for financial fraud risk assessment with publication number CN112419030A disclose a method, system and device for financial fraud risk assessment. The method comprises: obtaining financial statements and extracting financial indicator information; establishing a training sample set and a prediction sample set; converting the financial indicator information into qualitative indicators and quantitative indicators; predicting the prediction sample set with a random forest model to obtain a first data set of high-risk financial fraud; and / or predicting the prediction sample set with an adaptive enhancement model to obtain a second data set of high-risk financial fraud; and / or predicting the prediction sample set with a guided aggregation model to obtain a third data set of high-risk financial fraud; dividing the prediction sample set into sub-prediction sample sets; predicting the sub-prediction sample sets with indicator data to obtain a fourth data set of high-risk financial fraud; extracting data in the first, second, third and fourth data sets that exceeds a preset threshold to construct a high-risk financial fraud data set and generate a risk analysis report.

[0004] Internet consumer finance transactions are fast and frequent. Traditional evaluation methods are often based on historical data and have a long update cycle, which cannot reflect the current financial status and risk changes of users in real time. Traditional risk assessment models are mostly based on traditional financial business and lack adaptability to the unique scenarios and risk characteristics of internet consumer finance. The user group of internet consumer finance is extensive and the business scenarios are complex, with large differences in risk factors under different scenarios. The selection of indicators and the determination of weights often depend on the experience and subjective judgment of evaluators, resulting in a lack of objectivity and consistency in the evaluation results, making it difficult to accurately measure the financial risks of internet consumer finance. These are the problems we need to solve. SUMMARY

[0005] The present application aims to solve the problems in the background art by providing an internet financial risk assessment method based on multi-dimensional data fusion.

[0006] The technical scheme of the present application: an internet financial risk assessment method based on multi-dimensional data fusion, comprising the following steps:

[0007] S1, set internet data points, obtain user financial data, user information data and user timestamp, and construct a user financial tree through the user financial data and the user timestamp;

[0008] S2, monitor the user financial tree through the internet data points, obtain monitoring timestamp and financial monitoring data, analyze the monitoring timestamp and the financial monitoring data, and obtain user radius, financial orientation value and multi-dimensional financial fusion graph;

[0009] S3, analyze the user financial tree and the multi-dimensional financial fusion graph through the user timestamp, obtain the change time value, and obtain the risk trend value and the trend time value according to the multi-dimensional financial fusion graph and the change time value;

[0010] S4, predict through the risk trend value and the trend time value, obtain the predicted risk trend value and the predicted trend time value, assess the user's financial risk through the predicted risk trend value and the predicted trend time value, and obtain the assessment result.

[0011] Preferably, the process of setting internet data points, obtaining user financial data, user information data and user timestamp, and constructing a user financial tree through the user financial data and the user timestamp comprises:

[0012] The user financial data includes user behavior data and user credit data; the user behavior data includes consumption amount and consumption category; the user credit data includes credit score, loan times and overdue time length; the user information data includes income and user basic data;

[0013] A layer of nodes, a second layer of nodes and a third layer of nodes are set, and a user financial tree is obtained according to each user information data, corresponding user timestamp data, user financial data, a layer of nodes, a second layer of nodes and a third layer of nodes.

[0014] Preferably, the process of monitoring the user financial tree through the internet data points, obtaining monitoring timestamp and financial monitoring data, and analyzing the monitoring timestamp and the financial monitoring data to obtain the user radius comprises:

[0015] The monitoring timestamp and the financial monitoring data are analyzed through natural language technology to obtain financial change data, financial no-change data, no-change timestamp and change timestamp;

[0016] The financial change data includes a consumption change amount, a consumption change category, a credit change score, a loan change number, an overdue change duration, and a change income; a change channel is obtained according to the user financial data, the financial change data, a user timestamp, and a change timestamp; and a user radius is obtained according to the user timestamp and the change channel.

[0017] Preferably, the process of analyzing the monitoring timestamp and the financial monitoring data to obtain a financial orientation value includes:

[0018] A consumption orientation value D is obtained according to the consumption change amount and the consumption change category;

[0019] ;

[0020] wherein i is the consumption change number, a is the total consumption change number, d is a consumption change amount, and g is an income. is a consumption change amount and consumption amount difference, and g is an income. is a consumption change category. is a consumption change frequency.

[0021] A credit orientation value X is obtained according to the credit change score, the loan change number, and the overdue change duration.

[0022] ;

[0023] wherein j is the loan change number, b is the total loan change number, R is a credit score full score, and R is a credit score full score. is the credit change score after the jth loan change. is a credit change score weight. is the overdue change duration.

[0024] An income orientation value S is obtained according to the change income and the income.

[0025] ;

[0026] wherein z is a user expenditure, m is the change income number, c is the total change income number, sc is a user basic income, and cu is a user savings. is the weight coefficient of the mth income change. is the change income corresponding to the mth income change.

[0027] The financial orientation value is obtained according to the consumption orientation value, the credit orientation value, and the income orientation value.

[0028] Preferably, the process of analyzing the monitoring timestamp and the financial monitoring data to obtain a multi-dimensional financial fusion graph includes:

[0029] According to the user radius, the consumption guide value, the credit guide value and the income guide value, a user targeting circle, a consumption circle, a credit circle and an income circle are obtained, and according to the user targeting circle, the consumption circle, the credit circle and the income circle, a multi-dimensional financial fusion graph is obtained.

[0030] Preferably, the process of analyzing the user financial tree and the multi-dimensional financial fusion graph through the user timestamp to obtain a change time value, and according to the multi-dimensional financial fusion graph and the change time value, obtaining a risk trend value and a trend time value includes:

[0031] A change time threshold and a standard guide value are set, and a change time value is obtained according to the time corresponding to the user timestamp and the change timestamp linked by the change channel in the user financial tree;

[0032] The change time value, the change time threshold, the financial guide value and the standard guide value are analyzed, the financial guide value greater than or equal to the standard guide value corresponding to the change time value greater than or equal to the change time threshold is analyzed to obtain a trend factor, and according to the trend factor and the financial guide value corresponding thereto, a risk trend value is obtained, and according to the trend factor and the change time value, a trend time value is obtained.

[0033] Preferably, the process of predicting through the risk trend value and the trend time value to obtain a predicted risk trend value and a predicted trend time value, and the process of evaluating the financial risk of the user through the predicted risk trend value and the predicted trend time value includes:

[0034] A random forest model is constructed through a machine learning method, the risk trend value and the trend time value are predicted through the random forest model, and the predicted risk trend value and the predicted trend time value are output; a real-time time is obtained, a prediction time is obtained according to the real-time time and the trend time value, a risk trend threshold is set, when the predicted risk trend value is greater than or equal to the risk trend threshold, the real-time time to the prediction time corresponding to the financial risk is a risk rising period, the predicted risk trend value is sent to the Internet side for recording to obtain an evaluation result.

[0035] The application also discloses an Internet financial risk evaluation system based on multi-dimensional data fusion, comprising a management center, the management center is in communication connection with a data acquisition module, a data analysis module, a data processing module and a data evaluation module:

[0036] The data acquisition module is used for setting Internet data points, obtaining user financial data, user information data and user timestamps, and constructing a user financial tree through the user financial data and the user timestamps;

[0037] The data analysis module is used for monitoring the user financial tree through the internet data points, obtaining a monitoring timestamp and financial monitoring data, analyzing the monitoring timestamp and the financial monitoring data, and obtaining a user radius, a financial orientation value, and a multi-dimensional financial fusion graph.

[0038] The data processing module is used for analyzing the user financial tree and the multi-dimensional financial fusion graph through the user timestamp, obtaining a change time value, obtaining a risk trend value and a trend time value according to the multi-dimensional financial fusion graph and the change time value.

[0039] The data evaluation module is used for predicting the risk trend value and the trend time value, evaluating the financial risk of the user through the predicted risk trend value and the predicted trend time value, and obtaining an evaluation result.

[0040] Compared with the prior art, the above technical scheme of the present application has the following beneficial technical effects: by constructing the user financial tree, the management of the user financial data is optimized, and the efficiency of the user financial data analysis is improved; the monitoring timestamp and the financial monitoring data are obtained, the change of the financial related data is obtained and analyzed in time, the dynamic monitoring of the financial trend is realized, the timeliness of the data analysis is improved, and the dynamic environment change is adapted; the multi-dimensional data is analyzed by the user radius and the financial orientation value, the user is analyzed specifically, the accuracy of the financial risk identification is improved, and the financial risk state of each user is visually presented through the multi-dimensional financial fusion graph; the accuracy of the financial risk evaluation is improved with the help of the risk trend value and the trend time value; the risk prediction accuracy is improved through the predicted risk trend value and the predicted trend time value, the loan decision and the risk control strategy are optimized, and the efficiency of the decision support is improved. BRIEF DESCRIPTION OF DRAWINGS

[0041] Figure 1 A flowchart of an embodiment of the present application is shown. DETAILED DESCRIPTION

[0042] Embodiment one, as shown in the figure, the present application proposes an internet financial risk evaluation method based on multi-dimensional data fusion, which includes the following steps: Figure 1

[0043] S1, set the internet data points, obtain the user financial data, the user information data and the user timestamp, construct the user financial tree through the user financial data and the user timestamp;

[0044] S2, monitor the user financial tree through the internet data points, obtain a monitoring timestamp and financial monitoring data, analyze the monitoring timestamp and the financial monitoring data, and obtain a user radius, a financial orientation value, and a multi-dimensional financial fusion graph.​

[0045] S3, analyzing the user financial tree and the multi-dimensional financial fusion graph through the user timestamp to obtain a change time value, and obtaining a risk trend value and a trend time value according to the multi-dimensional financial fusion graph and the change time value;

[0046] S4, predicting through the risk trend value and the trend time value to obtain a predicted risk trend value and a predicted trend time value, and evaluating the financial risk of the user through the predicted risk trend value and the predicted trend time value to obtain an evaluation result.

[0047] It should be further explained that, in the specific implementation process, the user financial tree is constructed by setting the Internet data point, obtaining the user financial data, the user information data and the user timestamp, and through the user financial data and the user timestamp.

[0048] The user financial data includes user behavior data and user credit data; the user behavior data includes consumption amount and consumption category; the user credit data includes credit score, loan times and overdue time length; the user information data includes income and user basic data; and the user timestamp refers to a label related to the time sequence information of the Internet data point obtaining each user financial data.

[0049] It should be further explained that, in the specific implementation process, the user basic data includes user age, occupation, contact information and other basic personal information.

[0050] A layer of nodes, a second layer of nodes and a third layer of nodes are set, each user information data, corresponding user timestamp data and user financial data are stored in the first layer of nodes, the second layer of nodes and the third layer of nodes, and the first layer of nodes, the second layer of nodes and the third layer of nodes related to the same user are sequentially linked to obtain the user financial tree.

[0051] It should be further explained that, in the specific implementation process, there is a progressive relationship between the first layer of nodes, the second layer of nodes and the third layer of nodes, and the link between the first layer of nodes, the second layer of nodes and the third layer of nodes is the link between the layers.

[0052] It should be further explained that, in the specific implementation process, the user financial tree is monitored through the Internet data point to obtain a monitoring timestamp and financial monitoring data, and the user radius, the financial orientation value and the multi-dimensional financial fusion graph are obtained by analyzing the monitoring timestamp and the financial monitoring data.

[0053] The monitoring timestamp refers to the user timestamp corresponding to the financial monitoring data generated by the user financial tree monitored by the Internet data point.

[0054] Analyze monitoring timestamps and financial monitoring data through natural language technology to obtain financial change data, financial unchanged data, unchanged timestamps, and changed timestamps;

[0055] The financial change data refers to the data on changes in user information data and user financial data during the monitoring process, including consumption change amount, consumption change category, credit score change, loan change number, overdue change duration, and income change;

[0056] By monitoring the timestamp, the financial monitoring data is updated to the user's financial tree, and the user's financial data is matched with the financial change data, and the user's timestamp is matched with the change timestamp. A link channel is constructed and recorded as the change channel; the financial unchanged data is merged with the user's financial data;

[0057] Obtain the consumption orientation value D based on the consumption change amount and consumption change category;

[0058] ;

[0059] Among them, i is the number of consumption changes, a is the total number of consumption changes, is the difference between the consumption change amount and the consumption amount, g is income, is the consumption change category, is the frequency of consumption changes;

[0060] Obtain the credit guidance value X based on the credit change score, loan change times, and overdue change duration;

[0061] ;

[0062] Among them, j is the number of loan changes, b is the total number of loan changes, R is the maximum credit score, is the credit change score after the j-th loan change, is the weight of the credit change score, Change the duration of the overdue period;

[0063] According to the variable income and income, the income orientation value S is obtained;

[0064] ;

[0065] Among them, z is user expenditure, m is the number of variable income, and c is the total number of variable income. is the weight coefficient of the mth income change, is the variable income corresponding to the mth income change, sc is the user's basic income, and cu is the user's savings;

[0066] Count the total number of changed channels corresponding to the user's timestamp and record it as the user radius;

[0067] A user targeting circle is obtained by taking the value of the user radius as the diameter of a circle, and adding the values of the consumption orientation value, the credit orientation value, and the income orientation value to the value of the user radius, and taking the obtained calculation result as the radius to draw a circle with the center of the user targeting circle as the center. A consumption circle, a credit circle, and an income circle are obtained in turn. The centers of the user targeting circle, the consumption circle, the credit circle, and the income circle are overlapped to obtain a multi-dimensional financial fusion graph, and the consumption orientation value, the credit orientation value, and the income orientation value are recorded as a financial orientation value.

[0068] It needs to be further explained that, in the specific implementation process, the user financial tree and the multi-dimensional financial fusion graph are analyzed by the user timestamp to obtain a change time value. According to the multi-dimensional financial fusion graph and the change time value, a risk trend value and a trend time value are obtained.

[0069] A change time threshold and a standard orientation value are set. The time corresponding to the user timestamp and the change timestamp linked by the change channel in the user financial tree is calculated by difference to obtain a change time value.

[0070] The change time value, the change time threshold, the financial orientation value, and the standard orientation value are analyzed. The financial orientation value corresponding to the change time value greater than or equal to the change time threshold and greater than or equal to the standard orientation value is analyzed to obtain a trend factor. The trend factor and its corresponding financial orientation value are weighted to obtain a risk trend value. The trend factor and the change time value are multiplied to obtain a trend time value.

[0071] It needs to be further explained that, in the specific implementation process, the process of analyzing the financial orientation value corresponding to the change time value greater than or equal to the change time threshold and greater than or equal to the standard orientation value is as follows: if the financial orientation value corresponding to the change time value greater than or equal to the change time threshold and greater than or equal to the standard orientation value is the consumption orientation value, then the circles adjacent to the consumption circle and larger than the consumption circle and the circles smaller than the consumption circle are obtained and recorded as consumption adjacent one circle and consumption adjacent two circle. The consumption adjacent one circle and the consumption adjacent two circle are respectively calculated by the circle area difference with the consumption circle, and the obtained circle area difference is calculated by ratio to obtain a consumption trend factor. Similarly, if the financial orientation value corresponding to the change time value greater than or equal to the change time threshold and greater than or equal to the standard orientation value is the credit orientation value or the income orientation value, then a credit trend factor or an income trend factor is obtained. The obtained credit trend factor or income trend factor or consumption trend factor is collectively referred to as a trend factor.

[0072] It needs to be further explained that, in the specific implementation process, the risk trend value and the trend time value are predicted to obtain a predicted risk trend value and a predicted trend time value, the financial risk of the user is evaluated by the predicted risk trend value and the predicted trend time value to obtain an evaluation result, and the process is as follows:

[0073] A random forest model is constructed by a machine learning method, the risk trend value and the trend time value are predicted by the random forest model, the predicted risk trend value and the predicted trend time value are output, a real-time time is obtained, the real-time time is added to the trend time value to obtain a predicted time, a risk trend threshold is set, when the predicted risk trend value is greater than or equal to the risk trend threshold, the real-time time to the predicted time corresponding to the financial risk is a risk rising period, the predicted risk trend value is sent to the Internet side for recording to obtain an evaluation result, and when the predicted risk trend value is less than the risk trend threshold, the user financial data is continuously monitored.

[0074] In embodiment two, the application provides an Internet financial risk assessment system based on multi-dimensional data fusion, which is applied to the Internet financial risk assessment method based on multi-dimensional data fusion in embodiment one, and specifically comprises a management center, the management center is in communication connection with a data acquisition module, a data analysis module, a data processing module and a data evaluation module.

[0075] The data acquisition module is used for setting Internet data points, obtaining user financial data, user information data and user time stamps, and constructing a user financial tree through the user financial data and the user time stamps.

[0076] The data analysis module is used for monitoring the user financial tree through the Internet data points, obtaining a monitoring time stamp and financial monitoring data, analyzing the monitoring time stamp and the financial monitoring data to obtain a user radius, a financial orientation value and a multi-dimensional financial fusion graph.

[0077] The data processing module is used for analyzing the user financial tree and the multi-dimensional financial fusion graph through the user time stamp to obtain a change time value, and obtaining a risk trend value and a trend time value according to the multi-dimensional financial fusion graph and the change time value.

[0078] The data evaluation module is used for predicting the risk trend value and the trend time value to obtain a predicted risk trend value and a predicted trend time value, and evaluating the financial risk of the user by the predicted risk trend value and the predicted trend time value to obtain an evaluation result.

[0079] The embodiments of the application are described in detail above in combination with the drawings, but the application is not limited thereto, and various changes can be made within the knowledge of those skilled in the art without departing from the purpose of the application.

Claims

1. An Internet financial risk assessment method based on multi-dimensional data fusion, characterized in that, The method comprises the following steps: S1, setting an internet data point, obtaining user financial data, user information data and user timestamp, and constructing a user financial tree through the user financial data and the user timestamp; S2, monitoring the user financial tree through the internet data point, obtaining a monitoring timestamp and financial monitoring data, analyzing the monitoring timestamp and the financial monitoring data, and obtaining a user radius, a financial orientation value and a multi-dimensional financial fusion graph; S3, analyzing the user financial tree and the multi-dimensional financial fusion graph through the user timestamp, obtaining a change time value, and obtaining a risk trend value and a trend time value according to the multi-dimensional financial fusion graph and the change time value; S4, predicting through the risk trend value and the trend time value, obtaining a predicted risk trend value and a predicted trend time value, evaluating the financial risk of the user through the predicted risk trend value and the predicted trend time value, and obtaining an evaluation result; The process of setting an internet data point, obtaining user financial data, user information data and user timestamp, and constructing a user financial tree comprises: The user financial data comprises user behavior data and user credit data; the user behavior data comprises consumption amount and consumption category; the user credit data comprises credit score, loan times and overdue time length; the user information data comprises income and user basic data; A first layer node, a second layer node and a third layer node are set, and a user financial tree is obtained according to each user information data, corresponding user timestamp data, user financial data, the first layer node, the second layer node and the third layer node; The process of monitoring the user financial tree through the internet data point, obtaining a monitoring timestamp and financial monitoring data, analyzing the monitoring timestamp and the financial monitoring data, and obtaining a user radius and a financial orientation value comprises: The monitoring timestamp and the financial monitoring data are analyzed through natural language technology to obtain financial change data, financial non-change data, non-change timestamp and change timestamp; The financial change data comprises consumption change amount, consumption change category, credit change score, loan change times, overdue change time length and change income; a change channel is obtained according to the user financial data, the financial change data, the user timestamp and the change timestamp; and a user radius is obtained according to the user timestamp and the change channel; A consumption orientation value D is obtained according to the consumption change amount and the consumption change category; ； Wherein, i is the number of consumption changes, a is the total number of consumption changes, is the difference between the consumption change amount and the consumption amount, g is the income, is the consumption change category, is the consumption change frequency; A credit orientation value X is obtained according to the credit change score, the loan change times and the overdue change time length; ； wherein j is the number of loan changes, b is the total number of loan changes, R is the credit score full score, is the credit change score after the jth loan change, is the weight of the credit change score, is the overdue change duration; An income orientation value S is obtained according to the change income and the income; ； wherein z is the user expenditure, m is the number of variable income, c is the total number of variable income, is the weight coefficient of the mth variable income, is the mth variable income corresponding to the variable income, sc is the user's basic income, and cu is the user's savings. The financial orientation value is obtained according to the consumption orientation value, the credit orientation value and the income orientation value. 2.The Internet financial risk assessment method based on multi-dimensional data fusion according to claim 1, characterized in that, The process of analyzing the monitoring timestamp and the financial monitoring data to obtain a multi-dimensional financial fusion graph comprises: A user targeting circle, a consumption circle, a credit circle and an income circle are obtained according to the user radius, the consumption orientation value, the credit orientation value and the income orientation value, and a multi-dimensional financial fusion graph is obtained according to the user targeting circle, the consumption circle, the credit circle and the income circle. 3.The Internet financial risk assessment method based on multi-dimensional data fusion according to claim 2, characterized in that, The process of obtaining the change time value by analyzing the user financial tree and the multi-dimensional financial fusion graph through the user timestamp, obtaining the risk trend value and the trend time value according to the multi-dimensional financial fusion graph and the change time value comprises: setting the change time threshold and the standard guide value, obtaining the change time value according to the user timestamp and the time corresponding to the change timestamp linked by the change channel in the user financial tree; analyzing the change time value, the change time threshold, the financial guide value and the standard guide value, analyzing the financial guide value greater than or equal to the standard guide value corresponding to the change time value greater than or equal to the change time threshold, obtaining the trend factor, and obtaining the risk trend value according to the trend factor and the financial guide value corresponding thereto, and obtaining the trend time value according to the trend factor and the change time value. 4.The Internet financial risk assessment method based on multi-dimensional data fusion according to claim 3, characterized in that, The process of predicting the risk trend value and the trend time value through the risk trend value and the trend time value, and evaluating the financial risk of the user through the predicted risk trend value and the predicted trend time value comprises: constructing a random forest model through a machine learning method, predicting the risk trend value and the trend time value through the random forest model, outputting the predicted risk trend value and the predicted trend time value, obtaining the real-time time, obtaining the prediction time according to the real-time time and the trend time value, setting the risk trend threshold, when the predicted risk trend value is greater than or equal to the risk trend threshold, the real-time time to the prediction time corresponding to the financial risk is the risk rising period, and the predicted risk trend value is sent to the Internet side for recording to obtain the evaluation result.

5. An internet financial risk assessment system based on multi-dimensional data fusion, specifically applied to the internet financial risk assessment method based on multi-dimensional data fusion in any one of claims 1 to 4, comprising a management center, characterized in that, The management center is communicatively connected with a data acquisition module, a data analysis module, a data processing module and a data evaluation module: The data acquisition module is used to set the Internet data point, obtain the user financial data, the user information data and the user timestamp, and construct the user financial tree through the user financial data and the user timestamp; The data analysis module is used to monitor the user financial tree through the Internet data point, obtain the monitoring timestamp and the financial monitoring data, analyze the monitoring timestamp and the financial monitoring data, and obtain the user radius, the financial guide value and the multi-dimensional financial fusion graph; The data processing module is used to analyze the user financial tree and the multi-dimensional financial fusion graph through the user timestamp, obtain the change time value, and obtain the risk trend value and the trend time value according to the multi-dimensional financial fusion graph and the change time value; The data evaluation module is used to predict the risk trend value and the trend time value through the risk trend value and the trend time value, obtain the predicted risk trend value and the predicted trend time value, evaluate the financial risk of the user through the predicted risk trend value and the predicted trend time value, and obtain the evaluation result.

Images