Abnormal identification method and device of claim settlement case, computer equipment and storage medium
By obtaining and dividing the samples of claims cases, the target indicators and abnormal identification strategies are determined, and the problem of fraud in insurance claims is difficult to identify, and the accuracy and review efficiency of abnormal claims cases are improved.
Patent Information
- Application Number
- CN202411329105.1
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-09-23
- Publication Date
- 2025-01-24
AI Technical Summary
During the insurance claims process, fraud is diverse and difficult to identify, resulting in high audit costs and low accuracy. How to improve the accuracy of identification of abnormal claims cases has become an urgent problem.
By obtaining multiple preset claims case samples, including normal samples and abnormal samples, based on the preset indicator calculation rules, the candidate indicator information is determined, the samples are divided into sample subsets, the segmentation characteristics are calculated, the target indicators are determined, and based on the test Adjust the segmentation point of the case sample and determine the anomaly identification strategy.
The accuracy of the identification of abnormal claims is improved, the audit cost is reduced, and the abnormal identification strategy can be adjusted in real time to adapt to the actual situation.
Smart Images
Figure CN119359450A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of data analysis, and in particular to a method, device, computer equipment and storage medium for identifying anomalies in claims cases. Background Art
[0002] In order to reduce possible fraud in the insurance claims process, insurance companies usually pre-build relatively broad interception rules to screen out intercepted cases with abnormal characteristics, and then manually review the intercepted cases. Taking health insurance claims as an example, fraudulent behavior usually has some abnormal characteristics, such as abnormal time of accident, abnormal motivation for insurance, abnormal treatment information, abnormal claim materials, abnormal treatment institutions and abnormal claim application process; however, health insurance claims fraud types are diverse, including forged medical materials, false hospitalization, exaggeration of loss, fabrication of false accident causes, insurance after accident, and insurance with illness. It takes a lot of manpower and time to review intercepted cases, and the accuracy rate is not high. How to reduce the review cost of claim risk identification and improve the accuracy of the review has become an urgent problem to be solved. Summary of the invention
[0003] The main purpose of this application is to provide a method, device, equipment and computer storage medium for identifying abnormal claims cases, aiming to improve the accuracy of identifying abnormal claims cases.
[0004] In a first aspect, the present application provides a method for identifying anomalies in a claims case, the method comprising the following steps: Acquire a plurality of preset claim case samples, wherein the claim case samples include normal samples and abnormal samples; Based on a preset indicator calculation rule, determine the candidate indicator information corresponding to each of the claim case samples, and divide the claim case samples into at least one sample subset according to the candidate indicator information; Calculating the segmentation features of each of the sample subsets, and determining the target index from the candidate indexes according to the segmentation features; A first segmentation point of a sample subset corresponding to the target indicator is obtained, the first segmentation point is adjusted according to a test case sample to obtain a second segmentation point corresponding to the target indicator, and an anomaly identification strategy is determined according to the second segmentation point.
[0005] In a second aspect, the present application further provides a device for identifying anomalies in claims cases, the device for identifying anomalies in claims cases comprising: A sample acquisition module, used to acquire a plurality of preset claim case samples, wherein the claim case samples include normal samples and abnormal samples; A sample segmentation module, used to determine candidate indicator information corresponding to each of the claim case samples based on a preset indicator calculation rule, and segment the claim case samples into at least one sample subset according to the candidate indicator information; An indicator screening module, used to calculate the segmentation features of each of the sample subsets, and determine the target indicator from the candidate indicators according to the segmentation features; A strategy determination module is used to obtain a first segmentation point of a sample subset corresponding to the target indicator, adjust the first segmentation point according to a test case sample, obtain a second segmentation point corresponding to the target indicator, and determine an anomaly identification strategy according to the second segmentation point.
[0006] In a third aspect, the present application also provides a computer device, comprising a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein when the computer program is executed by the processor, the method for identifying anomalies in claims cases as described above is implemented.
[0007] In a fourth aspect, the present application also provides a computer-readable storage medium, on which a computer program is stored, wherein when the computer program is executed by a processor, the method for identifying anomalies in claims cases as described above is implemented.
[0008] The present application provides a method, device, equipment and computer storage medium for identifying abnormalities in claims cases. The present application obtains a plurality of preset claims case samples, wherein the claims case samples include normal samples and abnormal samples; based on preset indicator calculation rules, the candidate indicator information corresponding to each of the claims case samples is determined, and the claims case samples are divided into at least one sample subset according to the candidate indicator information; the segmentation features of each of the sample subsets are calculated, and the target indicator is determined from the candidate indicators according to the segmentation features; the first segmentation point of the sample subset corresponding to the target indicator is obtained, the first segmentation point is adjusted according to the test case sample, and the second segmentation point corresponding to the target indicator is obtained, and the abnormal identification strategy is determined according to the second segmentation point. An abnormal identification strategy that can distinguish between normal samples and abnormal samples is determined so as to identify claims cases with insurance fraud risks, and the segmentation point of the abnormal identification strategy is adjusted in real time according to the actual situation to improve the accuracy of abnormal claims case identification. BRIEF DESCRIPTION OF THE DRAWINGS
[0009] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.
[0010] Figure 1 A flowchart of a method for identifying anomalies in a claims case provided in one embodiment of the present application; Figure 2 A usage scenario diagram of a method for identifying anomalies in a claims case provided in one embodiment of the present application; Figure 3 A schematic block diagram of an abnormality identification device for a claim case provided in one embodiment of the present application; Figure 4 The present invention is a block diagram showing the structure of a computer device according to an embodiment of the present application. DETAILED DESCRIPTION
[0011] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0012] The flowcharts shown in the accompanying drawings are only examples and do not necessarily include all the contents and operations / steps, nor must they be executed in the order described. For example, some operations / steps may also be decomposed, combined or partially merged, so the actual execution order may change according to actual conditions.
[0013] Embodiments of the present application provide a method, apparatus, computer device, and computer-readable storage medium for identifying anomalies in claims cases.
[0014] In conjunction with the accompanying drawings, some embodiments of the present application are described in detail below. In the absence of conflict, the following embodiments and features in the embodiments can be combined with each other.
[0015] Please refer to Figure 1 , Figure 1 A flow chart of a method for identifying anomalies in a claims case provided for an embodiment of the present application. The method for identifying anomalies in a claims case can be used in a terminal or a server to determine the claims cases that can automatically identify the risk of insurance fraud. The terminal can be an electronic device such as a mobile phone, a tablet computer, a laptop computer, a desktop computer, a personal digital assistant, and a wearable device; the server can be an independent server, a server cluster, or 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 (CDNs), and big data and artificial intelligence platforms.
[0016] Please refer to Figure 2 , Figure 2 This is a usage scenario diagram provided by an embodiment of the present application. Figure 2 As shown, the corresponding candidate indicators are determined according to the claim case samples, the target indicators are screened out from the candidate indicators, and the segmentation points of the target indicators are adjusted to obtain the abnormality identification strategy. This method can be implemented in a terminal or a server, and is not limited here.
[0017] like Figure 1 As shown, the abnormality identification method for the claim case includes steps S101 to S104.
[0018] Step S101: Acquire a plurality of preset claim case samples, wherein the claim case samples include normal samples and abnormal samples.
[0019] In order to maintain market order and consumer rights, the identification of insurance fraud cases is crucial. However, in some types of health insurance, such as group health insurance claims, due to complex liability clauses and user situations, and low degree of data structuring, the accuracy of insurance fraud judgment in group health insurance claims is low and the risk penetration rate is high, resulting in the high claim rate of group health insurance, which may even exceed 100%.
[0020] In order to improve the accuracy of anomaly identification in claim cases, an embodiment of the present application provides an anomaly identification method for claim cases, which determines an anomaly identification strategy that can be applied to the current type of insurance through existing claim case samples in the database. The claim case samples in the database are classified into normal samples and abnormal samples based on document review and on-site investigation in the claims process. Among them, normal samples refer to claim cases in which claims are applied for according to actual circumstances, and abnormal samples refer to claim cases in which there are false reports, over-reporting and other insurance fraud behaviors in the claim application.
[0021] In some implementations, the obtaining of a plurality of preset claim case samples, wherein the claim case samples include normal samples and abnormal samples, comprises: A plurality of preset claim case samples are obtained, wherein a ratio of a target sample quantity of the abnormal samples to a total sample quantity of the claim case samples is less than a second preset ratio.
[0022] For example, in the process of classification according to the decision tree model, the balance of positive and negative samples is crucial to the accuracy of the model, and it is usually required that the ratio of the target sample number to the total sample number is about 5% for the best. However, in the abnormal identification method of claims cases provided in the embodiment of the present application, insurance fraud cases are a minority in the massive claims cases, and the number of samples is extremely unbalanced. The ratio of the target sample number of abnormal samples to the total sample number is small, about 0.5%, and the decision tree tends to split the majority class samples first during training, because doing so can increase information gain or reduce Gini impurity faster, causing the model to be biased towards the majority class, thereby ignoring the characteristics and rules of the minority class samples. Therefore, due to this characteristic of the claims case samples, it is necessary to adopt the abnormal identification method of claims cases provided in the embodiment of the present application to improve the accuracy of classification.
[0023] Step S102: based on a preset indicator calculation rule, determine the candidate indicator information corresponding to each of the claim case samples, and divide the claim case samples into at least one sample subset according to the candidate indicator information.
[0024] Exemplarily, the indicator calculation rule corresponding to each indicator is pre-set to calculate the candidate indicator information corresponding to the indicator. For example, the indicator may be the time interval between the insurance time and the effective time, and the indicator calculation rule may be to calculate the number of days between the date corresponding to the insurance time and the date corresponding to the effective time, so as to obtain the candidate indicator information, for example, 90 days.
[0025] Specifically, the types of candidate indicators may include policy terms data, such as underwriting type, insured amount, excluded diseases, number of insured units, gender of the insured, age, insurance time, effective time, etc.; it may also include current medical data, such as medical time, medical type, main diagnosis, secondary diagnosis, medical city, medical hospital, medical department, hospitalization days, bill amount, social security payment amount, time from first insurance, etc.; it may also include current claims data, such as claim amount, claim amount corresponding to each liability, whether to refuse, refused amount, reason for refusal, etc.; it may also include historical claims data, such as the number of historical claims of the insured, historical claim amount, whether to refuse, historical number of refused claims, whether to investigate, historical number of investigations, historical investigated diseases, time from the last investigation to the present, etc. Of course, it is not limited to this, and is not limited here.
[0026] Exemplarily, the claim case samples are segmented according to the candidate indicator information. The segmentation basis can be pre-set according to the actual situation, or the segmentation basis can be adaptively generated, which is not limited here. The segmentation basis can be one or more, and the claim case samples are segmented into at least one sample subset. For example, the claim case samples with a time interval between the insurance time and the effective time of less than or equal to 30 days and the claim case samples with a time interval between the insurance time and the effective time of more than 30 days are segmented into two sample subsets.
[0027] In some implementations, determining the candidate indicator information corresponding to each of the claim case samples based on a preset indicator calculation rule includes: Identify the claim case samples to obtain sample labels corresponding to the claim case samples, wherein the sample labels include at least one of: large-amount claim cases, serious illness claim cases, short-term claim cases, chronic illness claim cases, and first-time claim cases; According to the indicator calculation rules corresponding to the sample labels, the candidate indicator information corresponding to the claim case samples is calculated.
[0028] For example, for different types of claim case samples, the strategies for determining whether there are anomalies may be different. Therefore, the claim case samples need to be classified. Specifically, the claim case samples are divided into large claim cases, serious illness claim cases, short-term claim cases, chronic disease claim cases, and first-time claim cases.
[0029] Specifically, since large claims involve large amounts of money, policyholders may be attracted to defraud insurance money by fabricating or forging false accidents or exaggerating the extent of losses. In serious illness claims, common insurance fraud may include concealing one's condition when taking out insurance, that is, knowing that one has a certain disease before taking out insurance, but not informing the insurance company, and applying for claims after diagnosis. For short-term claims, that is, an insurance accident occurs and a claim is made within a short period of time after taking out insurance, there may be the possibility of deliberately fabricating an insurance accident or taking out insurance when it is known that an insurance accident is about to occur. For chronic disease claims, since chronic diseases usually require long-term treatment and medication, insurance fraud may include forging or exaggerating treatment costs and drug costs, or forging non-insured drug or treatment costs into insurance-covered costs for reimbursement. For first-time claims, since there is less historical data available for insurance companies to refer to, the authenticity of the case needs to be examined more carefully.
[0030] Therefore, for the above five types of claims cases, the target indicators used to judge abnormal cases may be different. Different indicators can be set in advance in the process of determining the abnormal identification strategy. The samples of each type of claims cases obtain the indicator calculation rules according to the corresponding sample labels, so as to obtain the candidate indicator information corresponding to the claims case samples. It can be seen that the types of candidate indicator information corresponding to the claims case samples with different sample labels may be different.
[0031] Step S103: Calculate the segmentation features of each of the sample subsets, and determine the target index from the candidate indexes according to the segmentation features.
[0032] Exemplarily, the segmentation feature is used to represent the classification ability of the candidate indicator information for the claim case sample. The simpler the type of the sample subset, the smaller the segmentation feature. Since the purpose of the anomaly identification strategy is to identify abnormal cases based on the target indicator, that is, it is necessary to gradually reduce the segmentation feature of the sample subset, the candidate indicator with a smaller segmentation feature is determined as the target indicator.
[0033] In some implementations, calculating the segmentation features of each of the sample subsets and determining the target index from the candidate indexes according to the segmentation features includes: The information gain ratio of each of the sample subsets is calculated, and a target indicator is determined from the candidate indicators according to the information gain ratio.
[0034] Exemplarily, the basis for selecting the target indicator may be the information gain ratio, but it is certainly not limited thereto and is not limited here.
[0035] Exemplarily, the information gain ratio balances the preference of information gain for the number of feature values by introducing a split information metric (e.g., introducing intrinsic information). Information gain and the Gini index tend to select features with more values, because the information entropy of the subset will be reduced more after such feature division, thereby obtaining a higher information gain. Compared with information gain and the Gini index, the information gain ratio adjusts this deviation by dividing by the intrinsic value of the attribute, so that features with fewer values are given more consideration in the decision tree division, which is more in line with the scenario of extremely unbalanced sample quantity in the embodiment of the present application.
[0036] In some implementations, calculating the information gain ratio of each of the sample subsets and determining the target indicator from the candidate indicators according to the information gain ratio includes: Calculate the information gain of the candidate indicator according to the information entropy of the claim case sample and the information entropy of each of the sample subsets; Calculating inherent information of the candidate indicators according to the proportion of each of the sample subsets; Determining the information gain ratio of the candidate indicator according to the ratio of the information gain to the inherent information; The candidate indicator whose information gain ratio is greater than the first preset ratio is determined as the target indicator.
[0037] Exemplarily, the information gain ratio is determined by the ratio of information gain to inherent information, that is, the information gain ratio , where information gain It indicates the degree of reduction of the information entropy of the data set D (i.e., the claim case sample) when the candidate indicator A is used as the classification basis. Therefore, it needs to be calculated based on the information entropy of the claim case sample and the information entropy of each sample subset. Specifically, information gain ,in, represents the information entropy of the data set D, Indicates that the candidate indicator A takes the value of v as a subset The number of samples, represents the number of samples in the dataset D, Representation Subset Information entropy. Intrinsic information ,in, represents the logarithm with base 2, and The meaning of is the same as that in the information gain calculation formula.
[0038] Exemplarily, the information gain ratio is adjusted by dividing the information gain by the intrinsic value of the feature. If feature A has many values, the intrinsic value is large and the information gain ratio is relatively small, thus avoiding the selection of features that have a large information gain simply because they have many values. On the contrary, if feature A has few values and a small intrinsic value, the information gain ratio may be relatively large, making such features more likely to be selected. This adjustment helps to select features that truly contribute to classification.
[0039] Step S104, obtaining a first segmentation point of the sample subset corresponding to the target indicator, adjusting the first segmentation point according to the test case sample to obtain a second segmentation point corresponding to the target indicator, and determining an anomaly identification strategy according to the second segmentation point.
[0040] For example, due to the complexity and variability of abnormal identification of claims cases and actual situations, the split point of the sample subset may change during the testing of the decision tree model, and it is necessary to capture the changes in time and make adjustments. For example, the first split point determined for the claim amount during the training process may be 50,000 yuan, and the reasonable split point of the claim amount during the training process may be greater than this value, so the second split point of the claim amount can be determined as 80,000 yuan.
[0041] In some implementations, adjusting the first segmentation point according to the test case sample to obtain a second segmentation point corresponding to the target indicator includes: Classifying the test case samples according to the first segmentation point corresponding to the target indicator, and determining a first coverage corresponding to the first segmentation point according to the classification result; When the first coverage is less than a preset coverage, the segmentation threshold of at least a portion of the target indicators is adjusted until the second segmentation point is obtained, and the second coverage corresponding to the second segmentation point meets a preset condition.
[0042] For example, in order to improve the accuracy of anomaly identification in claims cases, when the test case samples are segmented by the first segmentation point, it is necessary to cover as many abnormal cases as possible. The ability of the model to identify abnormal cases can be measured by coverage, and when the coverage does not meet the preset conditions, the segmentation threshold of the target indicator is adjusted to obtain the second segmentation point.
[0043] In some implementations, determining the first coverage corresponding to the first segmentation point according to the classification result includes: The first coverage is determined according to a ratio of a first number of abnormal samples in the classification result to a second number of abnormal samples in the test case sample.
[0044] Exemplarily, a first coverage for measuring the ability of a decision tree model to identify abnormal cases is obtained by dividing a first number of abnormal samples in the classification result hits by a second number of test case samples labeled as abnormal samples. When the first coverage is less than a preset coverage, the split point of the decision tree model is adjusted to improve the flexibility of the decision tree model.
[0045] The method for identifying abnormalities in claims cases provided in the above-mentioned embodiments obtains a plurality of preset claims case samples, wherein the claims case samples include normal samples and abnormal samples; based on preset indicator calculation rules, the candidate indicator information corresponding to each of the claims case samples is determined, and the claims case samples are divided into at least one sample subset according to the candidate indicator information; the segmentation features of each of the sample subsets are calculated, and the target indicator is determined from the candidate indicators according to the segmentation features; the first segmentation point of the sample subset corresponding to the target indicator is obtained, the first segmentation point is adjusted according to the test case sample, and the second segmentation point corresponding to the target indicator is obtained, and the abnormal identification strategy is determined according to the second segmentation point. An abnormal identification strategy that can distinguish between normal samples and abnormal samples is determined to identify claims cases with insurance fraud risks, and the segmentation point of the abnormal identification strategy is adjusted in real time according to the actual situation to improve the accuracy of abnormal claims case identification.
[0046] See also Figure 3 , Figure 3It is a schematic diagram of an abnormality identification device for a claims case provided in one embodiment of the present application. The abnormality identification device for a claims case can be configured in a server or a terminal to execute the aforementioned abnormality identification method for a claims case.
[0047] like Figure 3 As shown, the anomaly identification device for the claim case includes: a sample acquisition module 110, a sample segmentation module 120, an indicator screening module 130, and a strategy determination module 140.
[0048] The sample acquisition module 110 is used to acquire a plurality of preset claim case samples, wherein the claim case samples include normal samples and abnormal samples; The sample segmentation module 120 is used to determine the candidate indicator information corresponding to each of the claim case samples based on a preset indicator calculation rule, and segment the claim case samples into at least one sample subset according to the candidate indicator information; An indicator screening module 130, used to calculate the segmentation features of each of the sample subsets, and determine the target indicator from the candidate indicators according to the segmentation features; The strategy determination module 140 is used to obtain the first segmentation point of the sample subset corresponding to the target indicator, adjust the first segmentation point according to the test case sample, obtain the second segmentation point corresponding to the target indicator, and determine the anomaly identification strategy according to the second segmentation point.
[0049] In some implementations, the sample segmentation module 120, in the process of implementing the preset indicator calculation rule and determining the candidate indicator information corresponding to each of the claim case samples, is used to implement: Identify the claim case samples to obtain sample labels corresponding to the claim case samples, wherein the sample labels include at least one of: large-amount claim cases, serious illness claim cases, short-term claim cases, chronic illness claim cases, and first-time claim cases; According to the indicator calculation rules corresponding to the sample labels, the candidate indicator information corresponding to the claim case samples is calculated.
[0050] In some implementations, the indicator screening module 130, in the process of calculating the segmentation features of each of the sample subsets and determining the target indicator from the candidate indicators according to the segmentation features, is used to implement: The information gain ratio of each of the sample subsets is calculated, and a target indicator is determined from the candidate indicators according to the information gain ratio.
[0051] In some implementations, the indicator screening module 130, in the process of calculating the information gain ratio of each of the sample subsets and determining the target indicator from the candidate indicators according to the information gain ratio, is used to implement: Calculate the information gain of the candidate indicator according to the information entropy of the claim case sample and the information entropy of each of the sample subsets; Calculating inherent information of the candidate indicators according to the proportion of each of the sample subsets; Determining the information gain ratio of the candidate indicator according to the ratio of the information gain to the inherent information; The candidate indicator whose information gain ratio is greater than the first preset ratio is determined as the target indicator.
[0052] In some implementations, the strategy determination module 140, in the process of adjusting the first segmentation point according to the test case sample to obtain the second segmentation point corresponding to the target indicator, is used to implement: Classifying the test case samples according to the first segmentation point corresponding to the target indicator, and determining a first coverage corresponding to the first segmentation point according to the classification result; When the first coverage is less than a preset coverage, the segmentation threshold of at least a portion of the target indicators is adjusted until the second segmentation point is obtained, and the second coverage corresponding to the second segmentation point meets a preset condition.
[0053] In some implementations, the strategy determination module 104, in the process of determining the first coverage corresponding to the first segmentation point according to the classification result, is configured to implement: The first coverage is determined according to a ratio of a first number of abnormal samples in the classification result to a second number of abnormal samples in the test case sample.
[0054] In some embodiments, the abnormality identification device for claim cases is used to implement the process of obtaining a plurality of preset claim case samples, wherein the claim case samples include normal samples and abnormal samples: A plurality of preset claim case samples are obtained, wherein a ratio of a target sample quantity of the abnormal samples to a total sample quantity of the claim case samples is less than a second preset ratio.
[0055] It should be noted that those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the above-described devices and modules and units can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.
[0056] The method and apparatus of the present application can be used in many general or special computing system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.
[0057] Exemplarily, the above method and apparatus may be implemented in the form of a computer program. The computer program may be implemented in Figure 4 Runs on the computer device shown.
[0058] See also Figure 4 , Figure 4 A schematic block diagram of the structure of a computer device provided in an embodiment of the present application. The computer device may be a server or a terminal.
[0059] like Figure 4 As shown, the computer device includes a processor, a memory, and a network interface connected via a system bus, wherein the memory may include a storage medium and an internal memory.
[0060] The storage medium can store an operating system and a computer program. The computer program includes program instructions, and when the program instructions are executed, the processor can execute any abnormality identification method for a claim case.
[0061] The processor is used to provide computing and control capabilities and support the operation of the entire computer equipment.
[0062] The internal memory provides an environment for the operation of the computer program in the storage medium. When the computer program is executed by the processor, the processor can execute any abnormality identification method for a claim case.
[0063] The network interface is used for network communication, such as sending assigned tasks, etc. Those skilled in the art will understand that Figure 4The structure shown in the figure is only a block diagram of a part of the structure related to the solution of the present application, and does not constitute a limitation on the computer device to which the solution of the present application is applied. The specific computer device may include more or fewer components than those shown in the figure, or combine certain components, or have a different arrangement of components.
[0064] It should be understood that the processor may be a central processing unit (CPU), and the processor may also be other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Among them, the general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc.
[0065] In one embodiment, the processor is used to run a computer program stored in the memory to implement the following steps: Acquire a plurality of preset claim case samples, wherein the claim case samples include normal samples and abnormal samples; Based on a preset indicator calculation rule, determine the candidate indicator information corresponding to each of the claim case samples, and divide the claim case samples into at least one sample subset according to the candidate indicator information; Calculating the segmentation features of each of the sample subsets, and determining the target index from the candidate indexes according to the segmentation features; A first segmentation point of a sample subset corresponding to the target indicator is obtained, the first segmentation point is adjusted according to a test case sample to obtain a second segmentation point corresponding to the target indicator, and an anomaly identification strategy is determined according to the second segmentation point.
[0066] In one embodiment, when the processor implements the preset indicator calculation rule to determine the candidate indicator information corresponding to each of the claim case samples, it is used to implement: Identify the claim case samples to obtain sample labels corresponding to the claim case samples, wherein the sample labels include at least one of: large-amount claim cases, serious illness claim cases, short-term claim cases, chronic illness claim cases, and first-time claim cases; According to the indicator calculation rules corresponding to the sample labels, the candidate indicator information corresponding to the claim case samples is calculated.
[0067] In one embodiment, when the processor implements the step of calculating the segmentation features of each of the sample subsets and determining the target index from the candidate indexes according to the segmentation features, the processor implements: The information gain ratio of each of the sample subsets is calculated, and a target indicator is determined from the candidate indicators according to the information gain ratio.
[0068] In one embodiment, when the processor implements the step of calculating the information gain ratio of each of the sample subsets and determining the target indicator from the candidate indicators according to the information gain ratio, the processor is configured to implement: Calculate the information gain of the candidate indicator according to the information entropy of the claim case sample and the information entropy of each of the sample subsets; Calculating inherent information of the candidate indicators according to the proportion of each of the sample subsets; Determining the information gain ratio of the candidate indicator according to the ratio of the information gain to the inherent information; The candidate indicator whose information gain ratio is greater than the first preset ratio is determined as the target indicator.
[0069] In one embodiment, when the processor adjusts the first segmentation point according to the test case sample to obtain the second segmentation point corresponding to the target indicator, it is used to implement: Classifying the test case samples according to the first segmentation point corresponding to the target indicator, and determining a first coverage corresponding to the first segmentation point according to the classification result; When the first coverage is less than a preset coverage, the segmentation threshold of at least a portion of the target indicators is adjusted until the second segmentation point is obtained, and the second coverage corresponding to the second segmentation point meets a preset condition.
[0070] In one embodiment, when the processor determines the first coverage corresponding to the first segmentation point according to the classification result, it is configured to implement: The first coverage is determined according to a ratio of a first number of abnormal samples in the classification result to a second number of abnormal samples in the test case sample.
[0071] In one embodiment, when the processor implements the step of obtaining a plurality of preset claim case samples, wherein the claim case samples include normal samples and abnormal samples, the processor is configured to implement: A plurality of preset claim case samples are obtained, wherein a ratio of a target sample quantity of the abnormal samples to a total sample quantity of the claim case samples is less than a second preset ratio.
[0072] It should be noted that technical personnel in the relevant field can clearly understand that, for the convenience and conciseness of description, the specific working process of abnormal identification of claims cases described above can refer to the corresponding process in the aforementioned embodiment of the abnormal identification control method for claims cases, and will not be repeated here.
[0073] An embodiment of the present application also provides a computer-readable storage medium, on which a computer program is stored. The computer program includes program instructions. The method implemented when the program instructions are executed can refer to the various embodiments of the abnormality identification method for the claims case of the present application.
[0074] The computer-readable storage medium may be an internal storage unit of the computer device described in the foregoing embodiment, such as a hard disk or memory of the computer device. The computer-readable storage medium may also be an external storage device of the computer device, such as a plug-in hard disk, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc., equipped on the computer device.
[0075] It should be understood that the terms used in this application specification are only for the purpose of describing specific embodiments and are not intended to limit the application. As used in this application specification and the appended claims, unless the context clearly indicates otherwise, the singular forms "a", "an" and "the" are intended to include plural forms.
[0076] It should also be understood that the term "and / or" used in the specification of this application and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, including these combinations. It should be noted that, in this article, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or system including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or system. In the absence of further restrictions, an element defined by the sentence "including a..." does not exclude the presence of other identical elements in the process, method, article or system including the element.
[0077] The serial numbers of the embodiments of the present application are for description only and do not represent the advantages and disadvantages of the embodiments. The above description is only a specific implementation mode of the present application, but the protection scope of the present application is not limited thereto. Any technician familiar with the technical field can easily think of various equivalent modifications or replacements within the technical scope disclosed in the present application, and these modifications or replacements should be included in the protection scope of the present application. Therefore, the protection scope of the present application shall be based on the protection scope of the claims.
Claims
1. A method for identifying abnormalities in claims cases, characterized in that: The method comprises: Acquire a plurality of preset claim case samples, wherein the claim case samples include normal samples and abnormal samples; Based on a preset indicator calculation rule, determine the candidate indicator information corresponding to each of the claim case samples, and divide the claim case samples into at least one sample subset according to the candidate indicator information; Calculating the segmentation features of each of the sample subsets, and determining the target index from the candidate indexes according to the segmentation features; A first segmentation point of a sample subset corresponding to the target indicator is obtained, the first segmentation point is adjusted according to a test case sample to obtain a second segmentation point corresponding to the target indicator, and an anomaly identification strategy is determined according to the second segmentation point.
2. The method for identifying abnormalities in claims cases according to claim 1, characterized in that: The candidate indicator information corresponding to each of the claim case samples is determined based on the preset indicator calculation rules, including: Identify the claim case samples to obtain sample labels corresponding to the claim case samples, wherein the sample labels include at least one of: large-amount claim cases, serious illness claim cases, short-term claim cases, chronic illness claim cases, and first-time claim cases; According to the indicator calculation rules corresponding to the sample labels, the candidate indicator information corresponding to the claim case samples is calculated.
3. The method for identifying abnormalities in claims cases according to claim 1, characterized in that: The calculating the segmentation features of each of the sample subsets and determining the target index from the candidate indexes according to the segmentation features includes: The information gain ratio of each of the sample subsets is calculated, and a target indicator is determined from the candidate indicators according to the information gain ratio.
4. The method for identifying abnormalities in claims cases according to claim 3, characterized in that: The calculating the information gain ratio of each of the sample subsets and determining the target indicator from the candidate indicators according to the information gain ratio includes: Calculate the information gain of the candidate indicator according to the information entropy of the claim case sample and the information entropy of each of the sample subsets; Calculating inherent information of the candidate indicators according to the proportion of each of the sample subsets; Determining the information gain ratio of the candidate indicator according to the ratio of the information gain to the inherent information; The candidate indicator whose information gain ratio is greater than the first preset ratio is determined as the target indicator.
5. The method for identifying abnormalities in claims cases according to claim 1, characterized in that: The adjusting the first segmentation point according to the test case sample to obtain the second segmentation point corresponding to the target indicator includes: Classifying the test case samples according to the first segmentation point corresponding to the target indicator, and determining a first coverage corresponding to the first segmentation point according to the classification result; When the first coverage is less than a preset coverage, the segmentation threshold of at least a portion of the target indicators is adjusted until the second segmentation point is obtained, and the second coverage corresponding to the second segmentation point meets a preset condition.
6. The method for identifying abnormalities in claims cases according to claim 5, characterized in that: The determining, according to the classification result, a first coverage corresponding to the first segmentation point includes: The first coverage is determined according to a ratio of a first number of abnormal samples in the classification result to a second number of abnormal samples in the test case sample.
7. The method for identifying abnormalities in claims cases according to any one of claims 1 to 6, characterized in that: The obtaining of a plurality of preset claim case samples, wherein the claim case samples include normal samples and abnormal samples, comprises: A plurality of preset claim case samples are obtained, wherein a ratio of a target sample quantity of the abnormal samples to a total sample quantity of the claim case samples is less than a second preset ratio.
8. A device for identifying abnormalities in claims cases, characterized in that: The abnormality identification device for the claim case includes: A sample acquisition module, used to acquire a plurality of preset claim case samples, wherein the claim case samples include normal samples and abnormal samples; A sample segmentation module, used to determine candidate indicator information corresponding to each of the claim case samples based on a preset indicator calculation rule, and segment the claim case samples into at least one sample subset according to the candidate indicator information; An indicator screening module, used to calculate the segmentation features of each of the sample subsets, and determine the target indicator from the candidate indicators according to the segmentation features; A strategy determination module is used to obtain a first segmentation point of a sample subset corresponding to the target indicator, adjust the first segmentation point according to a test case sample, obtain a second segmentation point corresponding to the target indicator, and determine an anomaly identification strategy according to the second segmentation point.
9. A computer device, characterized in that: The computer device includes a processor, a memory, and a computer program stored in the memory and executable by the processor, wherein when the computer program is executed by the processor, the steps of the method for identifying anomalies in a claim case as described in any one of claims 1 to 7 are implemented.
10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, wherein when the computer program is executed by a processor, the steps of the method for identifying anomalies in a claims case according to any one of claims 1 to 7 are implemented.