Claims case diversion method, device, equipment and storage medium

Abstract

The present invention relates to artificial intelligence technology, and discloses a method, device, equipment and storage medium for diverting claims cases, the method comprising: training a model based on historical policy data, historical claims numbers and risk scores of historical claims policies to obtain a risk scoring model; obtaining the corresponding case information of the current insurance case, and inputting it into the risk scoring model to obtain a first risk score, wherein the case information includes scenario label information and insurance data; matching the corresponding decision rule from the rule engine based on the scenario label information, and inputting the insurance data into the decision rule to obtain a risk control adjustment factor; adjusting the first risk score based on the risk control adjustment factor to obtain a second risk score, and diverting the current insurance case according to the second risk score. After constructing the risk scoring model, the method cooperates with the rule engine to avoid the technical problem that the prediction accuracy of the risk model is low due to the small amount and latitude of case data when the claims scenario just appears.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence, and in particular to a method, device, equipment and storage medium for diverting claims cases. Background Art

[0002] With the development of the economy and the improvement of people's living standards, the insurance industry has been booming, especially the auto insurance market, which has attracted the influx of many insurance companies. At the same time, insurance claims have also attracted more and more attention from the public. Complaints about insurance claims are constantly coming, and claims are difficult to settle, and insurance fraud is also common. How to shorten the claim time limit, improve the claim service level, improve customer satisfaction, and enhance insurance brand loyalty have become the primary considerations of many insurance companies. Traditional insurance claims need to go through the stages of customer reporting, investigation, damage assessment, price and damage assessment, document collection, and payment and settlement. The cumbersome and inefficient process reduces the speed of payment, so it has been widely "criticized". However, with the in-depth application of digital transformation in the insurance industry, the claims process has also undergone tremendous changes. Relying on Internet technology, certain results have been achieved in the control of auto insurance claims, but there are still some problems in actual application, which cannot effectively adapt to the market environment. Loopholes in management processes and loopholes in the law have increased the moral risk of claims to a certain extent. Therefore, it is crucial to identify and control the risks of claims cases.

[0003] In the existing technology, risk identification of current claims cases after building a model through historical data can achieve better results, but the prerequisite is that the dimensions and data volume included in the historical data are required to be high. However, in the early stages of a claims scenario, there are relatively few historical high-risk cases and insufficient information, which can easily lead to a decrease in the accuracy of the constructed risk model and poor control over claims cases. Summary of the invention

[0004] The main purpose of the present invention is to solve the technical problem that the existing risk model has low prediction accuracy due to the small amount and latitude of case data when the claim scenario just occurs.

[0005] A first aspect of the present invention provides a method for diverting claims cases, comprising: obtaining historical policy data and historical claims data of historical claims policies, and historical risk scores of the historical claims policies; performing data preprocessing on the historical policy data and historical claims data, and performing model training based on the preprocessed historical policy data and historical claims data and the historical risk scores to obtain a risk scoring model; obtaining case information corresponding to a current insurance case, and inputting the current insurance case into the risk scoring model to obtain a first risk score for the current insurance case, wherein the case information includes scenario label information and insurance data; matching a decision rule corresponding to the current insurance case from a preset rule engine based on the scenario label information, and inputting the insurance data into the decision rule to obtain a risk control adjustment factor; adjusting the first risk score based on the risk control adjustment factor to obtain a second risk score, and diverting the current insurance case according to the second risk score.

[0006] Optionally, in a first implementation method of the first aspect of the present invention, the historical policy data and historical claims data are preprocessed, and a risk scoring model is obtained by model training based on the preprocessed historical policy data and historical claims data and the historical risk score, including: according to a preset classification dimension, the historical policy data and the historical claims data are divided into N feature groups, wherein N is a natural number greater than or equal to two; through a preset data mining method, M features are screened from the feature group, and the M features are used as risk factors for the historical claims policy, wherein M is a natural number greater than or equal to two; based on the LightGBM regression algorithm, the risk factor is used as an independent variable, and the historical risk score is used as a dependent variable to construct a risk scoring model.

[0007] Optionally, in a second implementation manner of the first aspect of the present invention, the data mining method includes correlation analysis and principal component analysis, and selecting M features from the feature group through a preset data mining method, and using the M features as risk factors for the historical claim settlement policy includes: performing a correlation analysis on all features in the feature group to obtain a first feature weight for all features; performing an attribution analysis on all features in the feature group to obtain a second feature weight for all features; and based on the first feature weight and the second feature weight, selecting M features from the feature group, and using the M features as risk factors for the historical claim settlement policy.

[0008] Optionally, in a third implementation method of the first aspect of the present invention, the risk scoring model is constructed based on the LightGBM regression algorithm using the risk factor as an independent variable and the historical risk score as a dependent variable, including: dividing the risk factor into a training set and a test set, and constructing an initial risk model based on the training set and the corresponding historical risk score according to the LightGBM regression algorithm; inputting the test set into the initial risk model for testing to obtain a predicted risk score of the initial risk model; calculating a preset loss function based on the historical risk score and the predicted risk score to obtain a loss function value; determining whether the loss function value is less than a preset threshold; if not, adjusting the model parameters of the initial risk model based on the loss function value, and re-training the model based on the training set and the corresponding initial risk model until the loss function value is less than the preset threshold; if so, using the initial risk model as the risk scoring model.

[0009] Optionally, in a fourth implementation method of the first aspect of the present invention, the decision rule corresponding to the current insurance case is matched from a preset rule engine based on the scenario label information, and the risk control adjustment factor is obtained by inputting the insurance data into the decision rule, including: determining the rule scenario corresponding to the current insurance case based on the scenario label information; calling the corresponding decision rule from a preset rule engine based on the rule scenario; matching the insurance data with the nodes in the RETE network corresponding to the decision rule input with the insurance data to obtain a node matching result; and calculating the risk adjustment factor corresponding to the current insurance case based on the node matching result.

[0010] Optionally, in a fifth implementation manner of the first aspect of the present invention, the first risk score is adjusted based on the risk control adjustment factor to obtain a second risk score, and the current insurance case is diverted according to the second risk score, including: based on the case volume of the historical claim policies, respectively calculating the weight coefficients of the risk control adjustment factor and the first risk score; based on the risk control adjustment factor and the first risk score and the corresponding weight coefficients, the first risk score is adjusted to obtain a second risk score; and the current insurance case is diverted according to the second risk score.

[0011] Optionally, in a sixth implementation method of the first aspect of the present invention, the case diversion of the current insurance case according to the second risk score includes: comparing the second risk score with a preset risk threshold, and judging the risk category of the current insurance case based on the comparison result; if the risk category of the current insurance case is a low-risk case, adding a process acceleration mark to the current insurance case, and the process acceleration mark is used to remind corresponding staff to speed up the claims process of the current insurance case; if the risk category of the current insurance case is a high-risk case, sending the current insurance case to corresponding staff.

[0012] The second aspect of the present invention provides a claims case diversion device, comprising: a historical data acquisition module, used to obtain historical policy data and historical claims data of historical claims policies, and historical risk scores of the historical claims policies; a model training module, used to preprocess the historical policy data and historical claims data, and perform model training based on the preprocessed historical policy data and historical claims data and the historical risk scores to obtain a risk scoring model; a scoring module, used to obtain the corresponding case information of the current insurance case, and input the current insurance case into the risk scoring model to obtain a first risk score of the current insurance case, wherein the case information includes scenario label information and insurance data; a rule engine module, used to match the decision rule corresponding to the current insurance case from a preset rule engine based on the scenario label information, and input the insurance data into the decision rule to obtain a risk control adjustment factor; a diversion module, used to adjust the first risk score based on the risk control adjustment factor to obtain a second risk score, and divert the current insurance case according to the second risk score.

[0013] Optionally, in a first implementation method of the second aspect of the present invention, the model training module specifically includes: a feature division unit, used to divide the historical policy data and the historical claims data into N feature groups according to a preset classification dimension, wherein N is a natural number greater than or equal to two; a data mining unit, used to screen M features from the feature group through a preset data mining method, and use the M features as risk factors for the historical claims policies, wherein M is a natural number greater than or equal to two; a model construction unit, used to construct a risk scoring model based on the LightGBM regression algorithm using the risk factor as an independent variable and the historical risk score as a dependent variable.

[0014] Optionally, in a second implementation method of the second aspect of the present invention, the data mining unit is specifically used to: perform correlation analysis on all features in the feature group to obtain first feature weights of all features; perform attribution analysis on all features in the feature group to obtain second feature weights of all features; based on the first feature weight and the second feature weight, screen M features from the feature group, and use the M features as risk factors for the historical claims policy.

[0015] Optionally, in a third implementation manner of the second aspect of the present invention, the model building unit is specifically used to: divide the risk factors into a training set and a test set, and build an initial risk model based on the training set and the corresponding historical risk score according to the LightGBM regression algorithm; input the test set into the initial risk model for testing to obtain a predicted risk score of the initial risk model; calculate a preset loss function based on the historical risk score and the predicted risk score to obtain a loss function value; determine whether the loss function value is less than a preset threshold; if not, adjust the model parameters of the initial risk model based on the loss function value, and re-train the model based on the training set and the corresponding initial risk model until the loss function value is less than the preset threshold; if so, use the initial risk model as a risk scoring model.

[0016] Optionally, in a fourth implementation method of the second aspect of the present invention, the rule engine module is specifically used to: determine the rule scenario corresponding to the current insurance case based on the scenario label information; call the corresponding decision rule from a preset rule engine based on the rule scenario; input the insurance data into a RETE network corresponding to the decision rule, match the insurance data with the nodes in the RETE network, and obtain a node matching result; based on the node matching result, calculate the risk adjustment factor corresponding to the current insurance case.

[0017] Optionally, in a fifth implementation manner of the second aspect of the present invention, the diversion module specifically includes: a calculation unit, used to calculate the risk control adjustment factor and the weight coefficient of the first risk score based on the case volume of the historical claim policies; an adjustment unit, used to adjust the first risk score based on the risk control adjustment factor and the first risk score and the corresponding weight coefficients to obtain a second risk score; and a case diversion unit, used to divert the current insurance case according to the second risk score.

[0018] Optionally, in a sixth implementation method of the second aspect of the present invention, the case diversion unit is specifically used to: compare the second risk score with a preset risk threshold, and determine the risk category of the current insurance case based on the comparison result; if the risk category of the current insurance case is a low-risk case, add a process acceleration mark to the current insurance case, and the process acceleration mark is used to remind corresponding staff to speed up the claims process of the current insurance case; if the risk category of the current insurance case is a high-risk case, send the current insurance case to the corresponding staff.

[0019] The third aspect of the present invention provides a claims case diversion device, comprising: a memory and at least one processor, wherein instructions are stored in the memory, and the memory and the at least one processor are interconnected via lines; the at least one processor calls the instructions in the memory so that the claims case diversion device executes the steps of the above-mentioned claims case diversion method.

[0020] A fourth aspect of the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores instructions, which, when executed on a computer, enable the computer to execute the steps of the above-mentioned method for diverting claims cases.

[0021] In the technical solution of the present invention, historical policy data and historical claims data of historical claim policies, as well as historical risk scores of historical claim policies are obtained; the historical policy data and historical claims data are preprocessed, and a risk scoring model is obtained by model training based on the preprocessed historical policy data, historical claims data and historical risk scores; the corresponding case information of the current insurance case is obtained, and the current insurance case is input into the risk scoring model to obtain a first risk score of the current insurance case, wherein the case information includes scenario label information and insurance data; based on the scenario label information, the decision rule corresponding to the current insurance case is matched from a preset rule engine, and the insurance data is input into the decision rule to obtain a risk control adjustment factor; the first risk score is adjusted based on the risk control adjustment factor to obtain a second risk score, and the current insurance case is case diverted according to the second risk score. This method aims at the risk scoring model of claims cases based on multi-perspective modeling and rule engine. It integrates claims data and policy data, constructs feature groups from multiple perspectives, and mines the main factors related to the risk size of claims cases through methods such as correlation analysis and principal component analysis. It uses the main factors as independent variables of the multivariate regression model, and adds the rule engine when constructing the risk scoring model to avoid the technical problem of low prediction accuracy of the risk model due to the small amount and latitude of case data when the claims scenario first occurs. BRIEF DESCRIPTION OF THE DRAWINGS

[0022] Figure 1It is a schematic diagram of a first embodiment of the method for diverting claims cases in an embodiment of the present invention;

[0023] Figure 2 It is a schematic diagram of a second embodiment of the method for diverting claims cases in an embodiment of the present invention;

[0024] Figure 3 It is a schematic diagram of a third embodiment of the method for diverting claims cases in an embodiment of the present invention;

[0025] Figure 4 This is a schematic diagram of a fourth embodiment of the method for diverting claims cases in an embodiment of the present invention;

[0026] Figure 5 A schematic diagram of an embodiment of a claim case diversion device in an embodiment of the present invention;

[0027] Figure 6 It is a schematic diagram of another embodiment of the claim case diversion device in the embodiment of the present invention;

[0028] Figure 7 It is a schematic diagram of an embodiment of a claims case diversion device in an embodiment of the present invention. DETAILED DESCRIPTION

[0029] In the technical solution of the present invention, historical policy data and historical claims data of historical claim policies, as well as historical risk scores of historical claim policies are obtained; the historical policy data and historical claims data are preprocessed, and a risk scoring model is obtained by model training based on the preprocessed historical policy data, historical claims data and historical risk scores; the corresponding case information of the current insurance case is obtained, and the current insurance case is input into the risk scoring model to obtain a first risk score of the current insurance case, wherein the case information includes scenario label information and insurance data; based on the scenario label information, the decision rule corresponding to the current insurance case is matched from a preset rule engine, and the insurance data is input into the decision rule to obtain a risk control adjustment factor; the first risk score is adjusted based on the risk control adjustment factor to obtain a second risk score, and the current insurance case is case diverted according to the second risk score. This method aims at the risk scoring model of claims cases based on multi-perspective modeling and rule engine. It integrates claims data and policy data, constructs feature groups from multiple perspectives, and mines the main factors related to the risk size of claims cases through methods such as correlation analysis and principal component analysis. It uses the main factors as independent variables of the multivariate regression model, and adds the rule engine when constructing the risk scoring model to avoid the technical problem of low prediction accuracy of the risk model due to the small amount and latitude of case data when the claims scenario first occurs.

[0030] The terms "first", "second", "third", "fourth", etc. (if any) in the specification and claims of the present invention and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchanged where appropriate, so that the embodiments described herein can be implemented in an order other than that illustrated or described herein. In addition, the terms "including" or "having" and any variations thereof are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device that includes a series of steps or units is not necessarily limited to those steps or units that are clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0031] For ease of understanding, the specific process of the embodiment of the present invention is described below. Figure 1 The first embodiment of the method for diverting claims cases in the embodiment of the present invention includes:

[0032] 101. Obtain historical policy data and historical claims data of historical claims-paid policies, as well as historical risk scores of historical claims-paid policies;

[0033] It is understandable that the execution subject of the present invention can be a claims case diversion device, or a terminal or a server, which is not limited here. The embodiment of the present invention is described by taking the server as the execution subject as an example. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms.

[0034] In this embodiment, model training is performed based on the claims data and policy data of historical claims-settled policies, wherein the historical claims-settled policies may be policies that have already been paid, for which risk assessment is performed after the claims are settled by the staff and a risk score is given, or the risk scores may be manually given by the staff for unpaid policies.

[0035] In this embodiment, for historical claim policies for which no compensation has been paid, the policy data and claim data therein can be updated to obtain daily update values ​​of the policy data and claim data, and then the daily update values ​​of each indicator data of the policy data and claim data are calculated according to the time dimension. The above-mentioned historical claim policies may refer to insurance categories formulated by insurance companies for different customer situations, and each insurance category contains multiple historical claim policies.

[0036] 102. Preprocess the historical policy data and the historical claims data, and perform model training based on the preprocessed historical policy data, historical claims data and historical risk scores to obtain a risk scoring model;

[0037] In this embodiment, data preprocessing is performed on historical claims data and policy data, mainly feature selection is performed on the historical claims data and policy data, and three feature groups are constructed from multiple dimensions around policy information, historical information and case information.

[0038] In this embodiment, based on three feature groups, the main factors affecting the risk of claims (mainly insurance fraud risk) are identified through correlation analysis, attribution analysis and principal component analysis (PCA). In this embodiment, the five dimensions of claim amount (excessive claim amount is more likely to face insurance fraud risk), claim reason (the more serious the claim reason, the less likely insurance fraud), employment status of claim applicant (unemployed people are more likely to exaggerate losses), education level of claim applicant (customers with higher education level are less likely to commit fraud), and historical credit of claim applicant (if a good credit rating is maintained, the possibility of fraud will also be small) can be selected as risk factors for claims.

[0039] In this embodiment, five risk factors are used as independent variables, and the risk score of the case is used as the dependent variable. The LightGBM regression algorithm is used to build a multivariate regression model. By continuously adjusting parameters (num_leaves, max_depth, learning_rate and other parameters), the mean square error (rmse) is reduced, the optimal model parameters are found, and the construction of the risk scoring model is completed.

[0040] 103. Obtain case information corresponding to the current insurance case, and input the current insurance case into a risk scoring model to obtain a first risk score of the current insurance case, wherein the case information includes scenario label information and insurance data;

[0041] In actual applications, when the application scenario corresponding to the historical claim policy is just constructed, or the corresponding financial product is just issued, there are fewer high-risk cases in the historical claim policy. The risk score output by the constructed risk scoring model cannot accurately assess the risk level of the current insurance case, and the output risk score needs to be adjusted.

[0042] 104. Match the decision rule corresponding to the current insurance case from the preset rule engine based on the scenario tag information, and input the insurance data into the decision rule to obtain the risk control adjustment factor;

[0043] In actual applications, the rule engine is a series of software systems that execute rules based on some algorithms. The rule engine integrates the set of facts and the set of rules passed into the system to trigger one or more business operations. Fact refers to the data required for rule execution, which is used to match the conditions of the rules, that is, the historical policy data and historical claims data in this proposal. Rules are usually implemented in business code in a declarative manner. In this embodiment, the rule engine can be at least one of Drools, OpenL Tablets, Easy Rules, and RuleBook.

[0044] In this embodiment, the rule engine can also be set in the model training process of the risk scoring model. The risk score output by the training set is corrected through the rule engine. Adding the rule engine to the existing model for appropriate correction can directly improve the accuracy of the risk scoring model. The rule engine contains rules formulated by claims adjusters and surveyors based on historical experience, which helps to improve the model.

[0045] 105. The first risk score is adjusted based on the risk control adjustment factor to obtain a second risk score, and the current insurance cases are diverted according to the second risk score.

[0046] In this embodiment, a risk adjustment formula can be preset, and the weight of the first risk score can be set based on the data volume and dimension of the historical policy data and the historical claims data. When the data volume and dimension of the historical policy data and the historical claims data are large, it means that the accuracy of the constructed risk scoring model is high, and only the rule engine needs to be slightly adjusted. Even when the accuracy of the risk scoring model reaches a certain level, the weight of the rule engine can be set to zero, and the weight of the first risk score can be set to a larger level. When the data volume and dimension of the historical policy data and the historical claims data are small, it means that the accuracy of the constructed risk scoring model is small, and the risk score mainly relies on the rule engine for slight adjustments. Even when the risk scoring model has no historical policy data and historical claims data for model training, the weight of the rule engine can be set to the maximum, and the weight of the first risk score can be set to a smaller level.

[0047] In this embodiment, after some new claims data and policy data are accumulated, these new data can be used to improve the risk scoring model to enhance the accuracy of case risk identification and the identification of high-risk customers.

[0048] In this embodiment, historical policy data and historical claims data of historical claim policies, as well as historical risk scores of historical claim policies are obtained; the historical policy data and historical claims data are preprocessed, and a risk scoring model is obtained based on the preprocessed historical policy data, historical claims data and historical risk scores. The corresponding case information of the current insurance case is obtained, and the current insurance case is input into the risk scoring model to obtain a first risk score for the current insurance case, wherein the case information includes scenario label information and insurance data; the decision rule corresponding to the current insurance case is matched from a preset rule engine based on the scenario label information, and the insurance data is input into the decision rule to obtain a risk control adjustment factor; the first risk score is adjusted based on the risk control adjustment factor to obtain a second risk score, and the current insurance case is diverted according to the second risk score. This method aims at the risk scoring model of claims cases based on multi-perspective modeling and rule engine. It integrates claims data and policy data, constructs feature groups from multiple perspectives, and mines the main factors related to the risk size of claims cases through methods such as correlation analysis and principal component analysis. It uses the main factors as independent variables of the multivariate regression model, and adds the rule engine when constructing the risk scoring model to avoid the technical problem of low prediction accuracy of the risk model due to the small amount and latitude of case data when the claims scenario first occurs.

[0049] See also Figure 2 The second embodiment of the method for diverting claims cases in the embodiment of the present invention includes:

[0050] 201. Obtain historical policy data and historical claims data of historical claims-paid policies, as well as historical risk scores of historical claims-paid policies;

[0051] 202. According to the preset classification dimension, the historical policy data and the historical claims data are divided into N feature groups, where N is a natural number greater than or equal to 2;

[0052] In this embodiment, feature selection is performed on historical claims data and policy data, and three feature groups are constructed from multiple dimensions around policy information, historical information, and case information, and multiple feature types are used to characterize the risk level of customers. For the policy information feature group, features such as the policy start / end date, vehicle information, policyholder information, premium information, and revision history are mainly considered; for the historical information feature group, features such as historical accident information, historical claims information, and personal credit information are mainly considered; for the case information feature group, features are constructed mainly from the time, location, cause, person / vehicle, etc. of the case.

[0053] 203. Perform correlation analysis on all features in the feature group to obtain first feature weights of all features;

[0054] 204. Perform attribution analysis on all features in the feature group to obtain second feature weights of all features;

[0055] 205. Based on the first feature weight and the second feature weight, M features are selected from the feature group, and the M features are used as risk factors of the historical claim insurance policy;

[0056] In this embodiment, the data mining method includes one or more of correlation analysis, attribution analysis and principal component analysis (PCA), which is not limited by the present invention. Among them, principal component analysis is when using statistical analysis methods to study multivariate topics. Too many variables will increase the complexity of the topic. People naturally hope that the number of variables is small and more information is obtained. In many cases, there is a certain correlation between variables. When there is a certain correlation between two variables, it can be interpreted that the information reflected by these two variables on this topic has a certain overlap. Principal component analysis is to delete redundant repeated variables (closely related variables) for all the variables originally proposed, and establish as few new variables as possible, so that these new variables are unrelated to each other, and these new variables keep the original information as much as possible in terms of reflecting the information of the topic.

[0057] In practical applications, the data mining method includes one or more of correlation analysis, attribution analysis and principal component analysis (PCA). In this embodiment, the data mining method includes correlation analysis and principal component analysis as an example, wherein correlation analysis refers to the analysis of two or more correlated variable elements to measure the degree of correlation between two factors. Correlation analysis can only be performed if there is a certain connection or probability between the correlated elements. Through correlation analysis and attribution analysis, the first feature weight and the second feature weight of each feature are calculated, and the features with the largest combined first feature weight and second feature weight are selected as the risk factors of the historical claims policy.

[0058] 206. Divide the risk factors into training sets and test sets, and build an initial risk model based on the training set and the corresponding historical risk scores according to the LightGBM regression algorithm;

[0059] In this embodiment, LightGBM is a gradient boosting framework that uses a tree-based learning algorithm. It is a version of XGB that improves performance, has similar accuracy to other GBMs and is 20 times faster to train. Regression refers to a statistical analysis method that studies the relationship between a set of random variables (Y1, Y2, ..., Yi) and another set of variables (X1, X2, ..., Xk), also known as multiple regression analysis. Usually Y1, Y2, ..., Yi are dependent variables, and X1, X2, ..., Xk are independent variables.

[0060] 207. Input the test set into the initial risk model for testing to obtain a predicted risk score of the initial risk model;

[0061] 208. Calculate a preset loss function based on the historical risk score and the predicted risk score to obtain a loss function value;

[0062] 209. Determine whether the loss function value is less than a preset threshold;

[0063] 210. If not, adjust the model parameters of the initial risk model based on the loss function value, and retrain the model based on the training set and the corresponding initial risk model until the loss function value is less than the preset threshold;

[0064] 211. If yes, the initial risk model is used as the risk scoring model;

[0065] In practical applications, the LightGBM regression algorithm is a new boosting framework released by Microsoft. The basic principle is the same as XGBoost. It uses a decision tree based on a learning algorithm, but it has been optimized on the framework (focusing on optimizing the training speed of the model). The most important thing is that LightGBM uses a histogram-based decision tree algorithm. The basic idea is to first discretize continuous floating-point eigenvalues ​​into k integers, and construct a histogram with a width of k. When traversing the data, the statistics are accumulated in the histogram based on the discretized values ​​as indexes. After traversing the data once, the histogram accumulates the required statistics. Then, according to the discrete values ​​of the histogram, the optimal segmentation point is traversed.

[0066] 212. Obtain case information corresponding to the current insurance case, and input the current insurance case into a risk scoring model to obtain a first risk score of the current insurance case, wherein the case information includes scenario label information and insurance data;

[0067] 213. Match the decision rule corresponding to the current insurance case from the preset rule engine based on the scenario tag information, and input the insurance data into the decision rule to obtain the risk control adjustment factor;

[0068] 214. The first risk score is adjusted based on the risk control adjustment factor to obtain a second risk score, and the current insurance cases are diverted according to the second risk score.

[0069] Based on the previous embodiment, this embodiment describes in detail the process of preprocessing the historical policy data and the historical claims data, and training the model based on the preprocessed historical policy data and the historical claims data and the historical risk score to obtain the risk scoring model. The historical policy data and the historical claims data are divided into N feature groups according to the preset classification dimensions; M features are selected from the feature group by the preset data mining method, and the M features are used as the risk factors of the historical claims policy; based on the LightGBM regression algorithm, the risk factor is used as an independent variable, and the historical risk score is used as a dependent variable to construct a risk scoring model. This embodiment integrates the claims data and the policy data, constructs a feature group from multiple perspectives, and mines the main factors related to the risk size of the claims case through methods such as correlation analysis and principal component analysis as the independent variables of the multivariate regression model. The independent variables selected here have certain mutual exclusivity and have a significant impact on the dependent variable (the risk score of the case), so that the model deviation caused by the correlation of the independent variables can be eliminated, and the accuracy of the risk scoring model is enhanced.

[0070] See also Figure 3 The third embodiment of the method for diverting claims cases in the embodiment of the present invention includes:

[0071] 301. Obtain historical policy data and historical claims data of historical claims-paid policies, as well as historical risk scores of historical claims-paid policies;

[0072] 302. Preprocess the historical policy data and the historical claims data, and perform model training based on the preprocessed historical policy data, historical claims data and historical risk scores to obtain a risk scoring model;

[0073] 303. Obtain case information corresponding to the current insurance case, and input the current insurance case into a risk scoring model to obtain a first risk score of the current insurance case, wherein the case information includes scenario label information and insurance data;

[0074] 304. Determine the rule scenario corresponding to the current insurance case based on the scenario label information;

[0075] 305. Calling corresponding decision rules from a preset rule engine based on the rule scenario;

[0076] 306. Input the insurance data into the RETE network corresponding to the decision rule, match the insurance data with the nodes in the RETE network, and obtain a node matching result;

[0077] 307. Based on the node matching results, calculate the risk adjustment factor corresponding to the current insurance case;

[0078] In practical applications, the Rete algorithm is an efficient pattern matching algorithm used to implement a production rule system. It avoids multiple evaluations of the same condition through caching. The Rete algorithm is a fast forward rule matching algorithm, and its matching speed is independent of the number of rules. Rete is Latin, and its English equivalent is net, which means network. The Rete algorithm forms a rete network for pattern matching, and uses two characteristics of rule-based systems, namely temporal redundancy and structural similarity, to improve the system pattern matching efficiency. In this embodiment, a rule engine is constructed with the Rete algorithm as the core, and the RETE network corresponding to the decision rule inputted with the insurance data is used to quickly obtain the node matching result. Based on the node matching result, the risk adjustment factor corresponding to the current insurance case is calculated. For example, the insurance data of the current insurance case is matched to a set of claims rules applicable to the current insurance case. The claims rules have corresponding risk formulas. Based on the risk formula, the risk adjustment factor corresponding to the current insurance case is calculated.

[0079] 308. Adjust the first risk score based on the risk control adjustment factor to obtain a second risk score, and divert the current insurance cases according to the second risk score.

[0080] Based on the previous embodiment, this embodiment describes in detail the process of matching the decision rule corresponding to the current insurance case from the preset rule engine based on the scenario label information, and inputting the insurance data into the decision rule to obtain the risk control adjustment factor, by determining the rule scenario corresponding to the current insurance case based on the scenario label information; calling the corresponding decision rule from the preset rule engine based on the rule scenario; matching the insurance data with the nodes in the RETE network corresponding to the decision rule input by the insurance data to obtain the node matching result; and calculating the risk adjustment factor corresponding to the current insurance case based on the node matching result. This embodiment takes into account the factors such as the small number of high-risk cases and insufficient information, and adds the rule engine to the risk scoring model when constructing it, thereby increasing the applicability of the model.

[0081] See also Figure 4 The fourth embodiment of the method for diverting claims cases in the embodiment of the present invention includes:

[0082] 401. Obtaining historical policy data and historical claims data of historical claims-paid policies, as well as historical risk scores of historical claims-paid policies;

[0083] 402. Preprocess the historical policy data and the historical claims data, and perform model training based on the preprocessed historical policy data, historical claims data and historical risk scores to obtain a risk scoring model;

[0084] 403. Obtain case information corresponding to the current insurance case, and input the current insurance case into a risk scoring model to obtain a first risk score of the current insurance case, wherein the case information includes scenario label information and insurance data;

[0085] 404. Based on the scenario tag information, a decision rule corresponding to the current insurance case is matched from a preset rule engine, and the insurance data is input into the decision rule to obtain a risk control adjustment factor;

[0086] 405. Based on the case volume of historical claims insurance policies, the weight coefficient of the risk control adjustment factor and the first risk score are calculated respectively;

[0087] 406. Based on the risk control adjustment factor and the first risk score and the corresponding weight coefficients, the first risk score is adjusted to obtain a second risk score;

[0088] In this embodiment, a risk adjustment formula can be preset, and the weight of the first risk score can be set based on the data volume and dimension of the historical policy data and the historical claims data. When the data volume and dimension of the historical policy data and the historical claims data are large, it means that the accuracy of the constructed risk scoring model is high, and only the rule engine needs to be slightly adjusted. Even when the accuracy of the risk scoring model reaches a certain level, the weight of the rule engine can be set to zero, and the weight of the first risk score can be set to a large value. When the data volume and dimension of the historical policy data and the historical claims data are small, it means that the accuracy of the constructed risk scoring model is small, and the risk score mainly relies on the rule engine for slight adjustments. Even when the risk scoring model has no historical policy data and historical claims data for model training, the weight of the rule engine can be set to the maximum, and the weight of the first risk score can be set to a small value. Based on the formula, the second risk score can be calculated.

[0089] 407. Compare the second risk score with a preset risk threshold, and determine the risk category of the current insurance case based on the comparison result;

[0090] In this embodiment, a risk threshold is set in advance. When the risk score of the current insurance case is greater than the risk threshold, the current insurance case is a high-risk case. When the risk score of the current insurance case is less than or equal to the risk threshold, the current insurance case is a low-risk case.

[0091] 408. If the risk category of the current insurance case is a low-risk case, the current insurance case is marked with a process acceleration mark, and the process acceleration mark is used to remind the corresponding staff to speed up the claim settlement process of the current insurance case;

[0092] 409. If the risk category of the current insurance case is a high-risk case, the current insurance case is sent to the corresponding staff.

[0093] In this embodiment, when a new claim case occurs, the risk scoring model is called in real time to divert the cases. Low-risk cases speed up the claims process, and high-risk cases are promptly notified to claims operators and investigators to help identify high-risk lists and reduce the possibility of fraudulent claims.

[0094] Based on the previous embodiment, this embodiment describes in detail the process of adjusting the first risk score based on the risk control adjustment factor to obtain the second risk score, and diverting the current insurance case according to the second risk score, by respectively calculating the weight coefficients of the risk control adjustment factor and the first risk score based on the case volume of the historical claim insurance policies; adjusting the first risk score based on the risk control adjustment factor and the first risk score and the corresponding weight coefficients to obtain the second risk score; and diverting the current insurance case according to the second risk score. This embodiment facilitates speeding up the claims process and improving claims efficiency by diverting insurance cases into low-risk cases or high-risk cases.

[0095] The above describes the method for diverting claims cases in the embodiment of the present invention. The following describes the device for diverting claims cases in the embodiment of the present invention. Figure 5 In one embodiment of the present invention, a device for diverting claims cases includes:

[0096] The historical data acquisition module 501 is used to acquire historical policy data and historical claims data of historical claims policies, as well as historical risk scores of the historical claims policies;

[0097] A model training module 502 is used to perform data preprocessing on the historical policy data and historical claims data, and to perform model training based on the preprocessed historical policy data and historical claims data and the historical risk scores to obtain a risk scoring model;

[0098] Scoring module 503, used to obtain case information corresponding to the current insurance case, and input the current insurance case into the risk scoring model to obtain a first risk score of the current insurance case, wherein the case information includes scenario label information and insurance data;

[0099] A rule engine module 504 is used to match the decision rule corresponding to the current insurance case from a preset rule engine based on the scenario tag information, and input the insurance data into the decision rule to obtain a risk control adjustment factor;

[0100] The diversion module 505 is used to adjust the first risk score based on the risk control adjustment factor to obtain a second risk score, and divert the current insurance case according to the second risk score.

[0101] In an embodiment of the present invention, the claim case diversion device runs the above-mentioned claim case diversion method, and the claim case diversion device obtains historical policy data and historical claims data of historical claim policies, as well as historical risk scores of historical claim policies; performs data preprocessing on the historical policy data and historical claims data, and performs model training based on the preprocessed historical policy data, historical claims data and historical risk scores to obtain a risk scoring model; obtains case information corresponding to the current insurance case, and inputs the current insurance case into the risk scoring model to obtain a first risk score for the current insurance case, wherein the case information includes scenario label information and insurance data; matches the decision rule corresponding to the current insurance case from a preset rule engine based on the scenario label information, and inputs the insurance data into the decision rule to obtain a risk control adjustment factor; adjusts the first risk score based on the risk control adjustment factor to obtain a second risk score, and diverts the current insurance case according to the second risk score. This method aims at the risk scoring model of claims cases based on multi-perspective modeling and rule engine. It integrates claims data and policy data, constructs feature groups from multiple perspectives, and mines the main factors related to the risk size of claims cases through methods such as correlation analysis and principal component analysis. It uses the main factors as independent variables of the multivariate regression model, and adds the rule engine when constructing the risk scoring model to avoid the technical problem of low prediction accuracy of the risk model due to the small amount and latitude of case data when the claims scenario first occurs.

[0102] See also Figure 6 The second embodiment of the claim case diversion device in the embodiment of the present invention includes:

[0103] The historical data acquisition module 501 is used to acquire historical policy data and historical claims data of historical claims policies, as well as historical risk scores of the historical claims policies;

[0104] A model training module 502 is used to perform data preprocessing on the historical policy data and historical claims data, and to perform model training based on the preprocessed historical policy data and historical claims data and the historical risk scores to obtain a risk scoring model;

[0105] Scoring module 503, used to obtain case information corresponding to the current insurance case, and input the current insurance case into the risk scoring model to obtain a first risk score of the current insurance case, wherein the case information includes scenario label information and insurance data;

[0106] A rule engine module 504 is used to match the decision rule corresponding to the current insurance case from a preset rule engine based on the scenario tag information, and input the insurance data into the decision rule to obtain a risk control adjustment factor;

[0107] The diversion module 505 is used to adjust the first risk score based on the risk control adjustment factor to obtain a second risk score, and divert the current insurance case according to the second risk score.

[0108] In this embodiment, the model training module 502 specifically includes: a feature division unit 5021, which is used to divide the historical policy data and the historical claims data into N feature groups according to a preset classification dimension, wherein N is a natural number greater than or equal to two; a data mining unit 5022, which is used to screen M features from the feature group through a preset data mining method, and use the M features as risk factors for the historical claims policies, wherein M is a natural number greater than or equal to two; a model construction unit 5023, which is used to construct a risk scoring model based on the LightGBM regression algorithm using the risk factor as an independent variable and the historical risk score as a dependent variable.

[0109] In this embodiment, the data mining unit 5022 is specifically used to: perform correlation analysis on all features in the feature group to obtain first feature weights of all features; perform attribution analysis on all features in the feature group to obtain second feature weights of all features; based on the first feature weight and the second feature weight, screen M features from the feature group, and use the M features as risk factors for the historical claim policy.

[0110] In this embodiment, the model building unit 5023 is specifically used to: divide the risk factors into a training set and a test set, and build an initial risk model based on the training set and the corresponding historical risk score according to the LightGBM regression algorithm; input the test set into the initial risk model for testing to obtain a predicted risk score of the initial risk model; calculate a preset loss function based on the historical risk score and the predicted risk score to obtain a loss function value; determine whether the loss function value is less than a preset threshold; if not, adjust the model parameters of the initial risk model based on the loss function value, and re-train the model based on the training set and the corresponding initial risk model until the loss function value is less than the preset threshold; if so, use the initial risk model as a risk scoring model.

[0111] In this embodiment, the rule engine module 504 is specifically used to: determine the rule scenario corresponding to the current insurance case based on the scenario label information; call the corresponding decision rule from the preset rule engine based on the rule scenario; input the insurance data into the RETE network corresponding to the decision rule, match the insurance data with the nodes in the RETE network to obtain a node matching result; and calculate the risk adjustment factor corresponding to the current insurance case based on the node matching result.

[0112] In this embodiment, the diversion module 505 specifically includes: a calculation unit 5051, which is used to calculate the risk control adjustment factor and the weight coefficient of the first risk score based on the case volume of the historical claim insurance policy; an adjustment unit 5052, which is used to adjust the first risk score based on the risk control adjustment factor and the first risk score and the corresponding weight coefficients to obtain a second risk score; a case diversion unit 5053, which is used to divert the current insurance case according to the second risk score.

[0113] In this embodiment, the case diversion unit 5053 is specifically used to: compare the second risk score with a preset risk threshold, and determine the risk category of the current insurance case based on the comparison result; if the risk category of the current insurance case is a low-risk case, add a process acceleration mark to the current insurance case, and the process acceleration mark is used to remind the corresponding staff to speed up the claims process of the current insurance case; if the risk category of the current insurance case is a high-risk case, send the current insurance case to the corresponding staff.

[0114] Based on the previous embodiment, this embodiment describes in detail the specific functions of each module and the unit composition of some modules. Based on the claims case risk scoring model of multi-perspective modeling and rule engine, the claims data and policy data are integrated, and feature groups are constructed from multiple perspectives. Through methods such as correlation analysis and principal component analysis, the main factors related to the risk size of claims cases are mined as independent variables of the multivariate regression model, and the rule engine is added when constructing the risk scoring model to avoid the technical problem of low prediction accuracy of the risk model due to the small amount and latitude of case data when the claims scenario just occurs.

[0115] above Figure 5 and Figure 6 The claims case diversion device in the embodiment of the present invention is described in detail from the perspective of modular functional entities. The claims case diversion device in the embodiment of the present invention is described in detail from the perspective of hardware processing.

[0116] Figure 7 : is a schematic diagram of the structure of a claim case diversion device provided by an embodiment of the present invention. The claim case diversion device 700 may have relatively large differences due to different configurations or performances, and may include one or more processors (central processing units, CPU) 710 (for example, one or more processors) and a memory 720, and one or more storage media 730 (for example, one or more mass storage devices) storing application programs 733 or data 732. Among them, the memory 720 and the storage medium 730 can be short-term storage or permanent storage. The program stored in the storage medium 730 may include one or more modules (not shown in the figure), and each module may include a series of instruction operations in the claim case diversion device 700. Furthermore, the processor 710 can be configured to communicate with the storage medium 730, and execute a series of instruction operations in the storage medium 730 on the claim case diversion device 700 to implement the steps of the above-mentioned claim case diversion method.

[0117] The claim case diversion device 700 may also include one or more power supplies 740, one or more wired or wireless network interfaces 750, one or more input and output interfaces 760, and / or one or more operating systems 731, such as Windows Serve, Mac OS X, Unix, Linux, FreeBSD, etc. It will be appreciated by those skilled in the art that Figure 7 The structure of the claims case diversion device shown does not constitute a limitation on the claims case diversion device provided in the present application, and may include more or fewer components than shown in the figure, or a combination of certain components, or a different arrangement of components.

[0118] The present invention also provides a computer-readable storage medium, which may be a non-volatile computer-readable storage medium or a volatile computer-readable storage medium. Instructions are stored in the computer-readable storage medium. When the instructions are executed on a computer, the computer executes the steps of the claim case diversion method.

[0119] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working process of the above-described system, device, or unit can refer to the corresponding process in the aforementioned method embodiment and will not be repeated here.

[0120] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention is essentially or the part that contributes to the prior art or all or part of the technical solution can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM), random access memory (RAM), disk or optical disk and other media that can store program codes.

[0121] As described above, the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit the same. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that the technical solutions described in the aforementioned embodiments may still be modified, or some of the technical features thereof may be replaced by equivalents. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for diverting claims cases, characterized in that: The claims diversion methods include: Obtaining historical policy data and historical claims data of historical claims-paid policies, as well as historical risk scores of the historical claims-paid policies; Preprocessing the historical policy data and the historical claims data, and performing model training based on the preprocessed historical policy data and the historical claims data and the historical risk scores to obtain a risk scoring model; Acquire case information corresponding to a current insurance case, and input the current insurance case into the risk scoring model to obtain a first risk score of the current insurance case, wherein the case information includes scenario label information and insurance data; Based on the scenario tag information, a decision rule corresponding to the current insurance case is matched from a preset rule engine, and the insurance data is input into the decision rule to obtain a risk control adjustment factor; Adjusting the first risk score based on the risk control adjustment factor to obtain a second risk score, and diverting the current insurance case according to the second risk score; Among them, the data preprocessing of the historical policy data and the historical claims data, and model training based on the preprocessed historical policy data and the historical claims data and the historical risk score to obtain a risk scoring model includes: dividing the historical policy data and the historical claims data into N feature groups according to a preset classification dimension, wherein N is a natural number greater than or equal to two; selecting M features from the feature group through a preset data mining method, and using the M features as risk factors for the historical claims policy, wherein M is a natural number greater than or equal to two; constructing a risk scoring model based on the LightGBM regression algorithm using the risk factor as an independent variable and the historical risk score as a dependent variable; The matching of the decision rule corresponding to the current insurance case from a preset rule engine based on the scenario label information, and inputting the insurance data into the decision rule to obtain the risk control adjustment factor includes: determining the rule scenario corresponding to the current insurance case based on the scenario label information; calling the corresponding decision rule from a preset rule engine based on the rule scenario; matching the insurance data with the nodes in the RETE network corresponding to the decision rule inputted with the insurance data to obtain a node matching result; and calculating the risk adjustment factor corresponding to the current insurance case based on the node matching result; The adjusting the first risk score based on the risk control adjustment factor to obtain a second risk score, and diverting the current insurance case according to the second risk score includes: calculating the weight coefficients of the risk control adjustment factor and the first risk score respectively based on the case volume of the historical claim insurance policies; adjusting the first risk score based on the risk control adjustment factor and the first risk score and the corresponding weight coefficients to obtain a second risk score; and diverting the current insurance case according to the second risk score.

2. The method for diverting claims according to claim 1, characterized in that: The data mining method includes correlation analysis and principal component analysis. The method of selecting M features from the feature group by the preset data mining method and using the M features as risk factors of the historical claim settlement policy includes: Performing correlation analysis on all features in the feature group to obtain first feature weights of all features; Performing attribution analysis on all features in the feature group to obtain second feature weights of all features; Based on the first feature weight and the second feature weight, M features are selected from the feature group, and the M features are used as risk factors for the historical claim settlement policy.

3. The method for diverting claims according to claim 2, characterized in that: The risk scoring model constructed by taking the risk factor as an independent variable and the historical risk score as a dependent variable based on the LightGBM regression algorithm includes: The risk factors are divided into a training set and a test set, and an initial risk model is constructed based on the training set and the corresponding historical risk scores according to the LightGBM regression algorithm; Inputting the test set into the initial risk model for testing to obtain a predicted risk score of the initial risk model; Calculating a preset loss function based on the historical risk score and the predicted risk score to obtain a loss function value; Determine whether the loss function value is less than a preset threshold; If not, adjusting the model parameters of the initial risk model based on the loss function value, and re-training the model based on the training set and the corresponding initial risk model until the loss function value is less than a preset threshold; If yes, the initial risk model is used as the risk scoring model.

4. The method for diverting claims according to claim 1, characterized in that: The case diversion of the current insurance case according to the second risk score comprises: Comparing the second risk score with a preset risk threshold, and determining the risk category of the current insurance case based on the comparison result; If the risk category of the current insurance case is a low-risk case, a process acceleration mark is added to the current insurance case, and the process acceleration mark is used to remind the corresponding staff to speed up the claim settlement process of the current insurance case; If the risk category of the current insurance case is a high-risk case, the current insurance case is sent to a corresponding staff member.

5. A claim case diversion device, characterized in that: The claim case diversion device comprises: A historical data acquisition module, used to acquire historical policy data and historical claims data of historical claims policies, as well as historical risk scores of the historical claims policies; A model training module, used for preprocessing the historical policy data and the historical claims data, and performing model training based on the preprocessed historical policy data and the historical claims data and the historical risk scores to obtain a risk scoring model; A scoring module, used to obtain case information corresponding to a current insurance case, and input the current insurance case into the risk scoring model to obtain a first risk score of the current insurance case, wherein the case information includes scenario label information and insurance data; A rule engine module, used to match the decision rule corresponding to the current insurance case from a preset rule engine based on the scenario tag information, and input the insurance data into the decision rule to obtain a risk control adjustment factor; a diversion module, configured to adjust the first risk score based on the risk control adjustment factor to obtain a second risk score, and divert the current insurance case according to the second risk score; Among them, the data preprocessing of the historical policy data and the historical claims data, and model training based on the preprocessed historical policy data and the historical claims data and the historical risk score to obtain a risk scoring model includes: dividing the historical policy data and the historical claims data into N feature groups according to a preset classification dimension, wherein N is a natural number greater than or equal to two; selecting M features from the feature group through a preset data mining method, and using the M features as risk factors for the historical claims policy, wherein M is a natural number greater than or equal to two; constructing a risk scoring model based on the LightGBM regression algorithm using the risk factor as an independent variable and the historical risk score as a dependent variable; The matching of the decision rule corresponding to the current insurance case from a preset rule engine based on the scenario label information, and inputting the insurance data into the decision rule to obtain the risk control adjustment factor includes: determining the rule scenario corresponding to the current insurance case based on the scenario label information; calling the corresponding decision rule from a preset rule engine based on the rule scenario; matching the insurance data with the nodes in the RETE network corresponding to the decision rule inputted with the insurance data to obtain a node matching result; and calculating the risk adjustment factor corresponding to the current insurance case based on the node matching result; The adjusting the first risk score based on the risk control adjustment factor to obtain a second risk score, and diverting the current insurance case according to the second risk score includes: calculating the weight coefficients of the risk control adjustment factor and the first risk score respectively based on the case volume of the historical claim insurance policies; adjusting the first risk score based on the risk control adjustment factor and the first risk score and the corresponding weight coefficients to obtain a second risk score; and diverting the current insurance case according to the second risk score.

6. A claim case diversion device, characterized in that: The claim case diversion device comprises: a memory and at least one processor, the memory stores instructions, and the memory and the at least one processor are interconnected via a line; The at least one processor calls the instructions in the memory to enable the claim case diversion device to perform the steps of the claim case diversion method as described in any one of claims 1-4.

7. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the claim case diversion method as described in any one of claims 1 to 4 are implemented.

Images