KR102149542B1 - Device and method for detecting new fraudulent claims - Google Patents

Abstract

The present invention relates to a new fraudulent claim detection device and a method thereof. The new fraudulent claim detection device can comprise: a data preprocessing part that preprocesses insurance claim details data to derive and refine feature variables required for artificial intelligence learning; a model construction part that builds a normal autoencoder model that learns normal feature variables derived from normal insurance claim details data and an abnormal autoencoder model that learns abnormal feature variables derived from abnormal insurance claim details data; a new pattern detection part that detects a new pattern of fraudulent claims from an error value of each of the normal autoencoder model and the abnormal autoencoder model; and a new fraudulent claim determination part that determines new fraudulent insurance claims based on a fraudulent claim detection model that learns the new pattern.

Description

New unfair claim detection device and method {DEVICE AND METHOD FOR DETECTING NEW FRAUDULENT CLAIMS}

The present application relates to a novel apparatus and method for detecting unfair claims.

Existing unfair claims for insurance claims are mostly detected based on existing business rules to detect unfair claims and prevent insurance fraud by deriving rules based on expert experience and knowledge.

However, the expert's rule-based unfair claim method is often inefficient in terms of time and economy. Human resources are limited to detect unfair claims for insurance claims based on expert research, and time and economic constraints are imposed on processing all data in the era of big data.

In addition, insurance companies cannot neglect insurance money leaks without taking any action against insurance fraud, nor can they spend infinite investigation costs for the purpose of fully revealing insurance fraud. Accordingly, insurance companies need to determine the level of investigation efforts in an appropriate manner not to spend too much investigation cost while reducing the amount of insurance leakage from insurance fraud. Therefore, the existing insurance fraud prevention system needs a new method and a method capable of detecting new fraud patterns.

The technology behind the present application is disclosed in Korean Patent Application Publication No. 10-0862181.

The present application is to solve the problems of the prior art described above, an object of the present invention to provide a new unfair claim detection apparatus and method for detecting a new pattern of unfair claims and unfair claims of insurance money.

The present application is to solve the problems of the prior art described above, a new unfair claim detection device and method capable of learning unfair claim data to analyze a new unfair claim pattern, and to determine whether or not a new insurance claim is unfairly claimed. It aims to provide.

However, the technical problems to be achieved by the embodiments of the present application are not limited to the technical problems as described above, and other technical problems may exist.

As a technical means for achieving the above technical problem, a new unfair claim detection device according to an embodiment of the present application preprocesses insurance claim details data, and a data preprocessing unit that derives and refines characteristic variables required for artificial intelligence learning, A model construction unit that builds a normal autoencoder model that learns the normal feature variable derived from the normal insurance claim details data and the abnormal autoencoder model that learns the abnormal feature variable derived from the abnormal insurance claim detail data, the normal autoencoder model, and A new pattern detection unit that detects a new pattern of unfair claims from the error values of each of the abnormal autoencoder models, and a new unfair claim determination unit that determines unfair claims of new insurance claims based on the unfair claims detection model that learns the new pattern. can do.

According to an embodiment of the present application, further comprising a database for storing insurance claim detail data including the unstructured data, wherein the insurance claim detail data is at least one of billing data, contract data, payment data, insurance planner data, and customer data. It can contain one.

According to the exemplary embodiment of the present application, the model building unit may construct the normal autoencoder model and the abnormal autoencoder model based on an autoencoder algorithm based on unsupervised learning.

According to an exemplary embodiment of the present disclosure, the new pattern detection unit sets a threshold section for detecting a new pattern, and the threshold section may include normal, suspect, and abnormal.

According to an exemplary embodiment of the present disclosure, the new pattern detection unit may classify and label the threshold section based on a difference between an error value of a normal autoencoder model and an error value of an abnormal autoencoder model.

According to an exemplary embodiment of the present disclosure, the new pattern detection unit may detect the new pattern from learning data associated with an error value labeled as the suspect group.

According to an exemplary embodiment of the present application, the new unfair claim determination unit may determine an unfair claim of a new insurance claim based on the new pattern learning model and the new pattern based on a supervised learning-based classification algorithm.

In the method of detecting a new unfair claim by a new unfair claim detection device according to an embodiment of the present application, the step of deriving and refining the characteristic variables required for artificial intelligence learning by preprocessing the insurance claim details data, and deriving from the normal insurance claim details data Constructing a normal autoencoder model for learning a normal feature variable and an abnormal autoencoder model for learning an abnormal feature variable derived from the abnormal insurance claim details data, the error value of each of the normal autoencoder model and the abnormal autoencoder model The method may include detecting a new pattern of unfair claims from and determining an unfair claim of a new insurance claim based on an unfair claim detection model for learning the new pattern.

According to an embodiment of the present application, further comprising the step of storing insurance claim details data including the unstructured data, wherein the insurance claim details data is at least one of claim data, contract data, payment data, insurance agent data, and customer data. It can contain one.

According to an exemplary embodiment of the present application, in the step of constructing the model, the normal autoencoder model and the abnormal autoencoder model may be constructed based on an autoencoder algorithm based on unsupervised learning.

According to an exemplary embodiment of the present disclosure, in the detecting of the new pattern, a threshold section for detecting the new pattern is set, and the threshold section may include normal, suspect, and abnormal.

According to an exemplary embodiment of the present disclosure, the detecting of the new pattern may divide and label the threshold section based on a difference between an error value of a normal autoencoder model and an error value of an abnormal autoencoder model.

According to an exemplary embodiment of the present disclosure, the detecting of the new pattern may detect the new pattern from learning data associated with an error value labeled as the suspect group.

According to an embodiment of the present application, the step of determining the unfair claim of the new insurance claim includes determining the unfair claim of the new insurance claim based on the new pattern learning model and the new pattern based on a supervised learning-based classification algorithm. I can.

The above-described problem solving means are merely exemplary and should not be construed as limiting the present application. In addition to the above-described exemplary embodiments, additional embodiments may exist in the drawings and detailed description of the invention.

According to the above-described problem solving means of the present application, it is possible to provide a new unfair claim detection apparatus and method for detecting an unfair claim pattern of insurance money and a new pattern of unfair claim.

According to the above-described problem solving means of the present application, it is possible to provide a new unfair claim detection apparatus and method capable of learning unfair claim data to analyze a new unfair claim pattern, and to determine whether or not a new insurance claim is unfairly claimed. .

1 is a diagram showing the configuration of a new unfair claim detection apparatus according to an embodiment of the present application.
2 is a diagram showing an example of building an autoencoder model of a new unfair claim detection apparatus according to an embodiment of the present application.
3 is a diagram illustrating a comparison of a normal error value and an abnormal error value of a new unfair claim detection apparatus according to an embodiment of the present application.
4 is a diagram illustrating an example of labeling of a new unfair claim detection device according to an embodiment of the present application.
5 is a diagram illustrating an example of detection of a new unfair pattern by a new unfair claim detection apparatus according to an embodiment of the present application.
6 is a diagram illustrating a flow of a method for detecting a new unfair claim according to an embodiment of the present application.

Hereinafter, exemplary embodiments of the present application will be described in detail with reference to the accompanying drawings so that those of ordinary skill in the art may easily implement the present application. However, the present application may be implemented in various different forms and is not limited to the embodiments described herein. In addition, in the drawings, parts not related to the description are omitted in order to clearly describe the present application, and similar reference numerals are attached to similar parts throughout the specification.

Throughout this specification, when a part is said to be "connected" with another part, this includes not only the case that it is "directly connected", but also the case that it is "electrically connected" with another element interposed therebetween. do.

Throughout this specification, when a member is positioned "on", "upper", "upper", "under", "lower", and "lower" of another member, this means that a member is located on another member. It includes not only the case where they are in contact but also the case where another member exists between the two members.

Throughout the specification of the present application, when a certain part "includes" a certain component, it means that other components may be further included rather than excluding other components unless specifically stated to the contrary.

1 is a diagram showing the configuration of a new unfair claim detection apparatus according to an embodiment of the present application.

Referring to FIG. 1, a new unfair claim detection apparatus 100 includes a data preprocessing unit 110, a model construction unit 120, a new pattern detection unit 130, a new unfair claim determination unit 140, and a database 150. Can include. The data preprocessor 110 may pre-process the insurance claim details data to derive and refine characteristic variables required for artificial intelligence learning.

Although it is self-evident that the following description includes not only unfair claims for insurance money, but also unfair claims such as subsidies, support funds, and security deposits, the explanation focuses on unfair claims for insurance money for convenience of explanation. The insurance claim detail data may include at least one of claims data for claiming insurance, insurance or subsidy related contract data, insurance payment data, insurance agent data, customer data, and insurance fraud detection result data.

The data pre-processing unit 110 may pre-process and format insurance claim details data including unstructured data, and may extract characteristic variables from the standardized insurance claim details data. For example, in the case of customer data, the standardization of insurance claim data refers to quantifying the income level of the customer, the number of hospitals visited by the customer, and the number of disease reasons the customer has applied for. In addition, the characteristic variable derived by the data preprocessor 110 is a variable that can have a numerical value related to the insurance claim. For example, the customer ID number, whether the insurance fraudster, the number of duplicate applications under the same name, The number of contracts signed, the maximum number of contracts per day, the number of securities approved for payment, the number of securities applied for payment, the number of securities signed by the customer, the number of insurance claims, the amount of credit rating change, the type of contract signed, the total number of visits to the hospital. The number of reasons for illness requested by the customer, the number of doctors the customer has met, the number of hospitals that the customer has visited, the total number of valid admission/outpatient days, the number of treatment subjects, the customer's income level, the number of FP (Financial Planner) changes, the number of loss treatment, It may include at least one of the number of fraudulent FP contracts.

The database 150 may store insurance claim detail data including unstructured data. The data preprocessor 110 may calculate a characteristic variable from the insurance claim detail data stored in the database. For example, in the database 150, insurance claim details data classified as normal claims (hereinafter, normal insurance claims details data) and insurance claim details data classified as abnormal claims (abnormal claims) (hereinafter, abnormal insurance claim details data). Can be separated and stored. The data preprocessor 110 may calculate characteristic variables from each of the normal insurance claim details data and the abnormal insurance claim details data.

The model building unit 120 may build a normal autoencoder model that learns a normal characteristic variable derived from the normal insurance claim details data and an abnormal autoencoder model that learns an abnormal characteristic variable derived from the abnormal insurance claim details data. According to an embodiment of the present application, the data preprocessing unit 110 may distinguish between normal and abnormal insurance claim details data in the process of calculating characteristic variables through pre-processing of insurance claim details data that are not classified into normal and abnormal. I can. Specifically, the data preprocessing unit 110 standardizes each element (claim data, contract data, etc.) included in the above-described insurance claim detail data, and makes a corresponding insurance claim according to whether or not a preset threshold value is exceeded for each of the elements. It is possible to distinguish whether the historical data is normal insurance claim history data or abnormal insurance claim history data. In addition, the model building unit 120 may learn characteristic variables derived from the normal insurance claim details data and the abnormal insurance claim details data separated by the data preprocessing unit 110.

The model building unit 120 may build a normal autoencoder model and an abnormal autoencoder model based on an unsupervised learning-based autoencoder algorithm. The main feature of the autoencoder algorithm is to learn each layer of a neural network step by step so that the final output reproduces the initial input. In other words, it means to accept an input value and produce an output of the same size. At this time. The number of nodes in the hidden layer is smaller than that of the input node and the output node. In other words, it means that the data is compressed. As data is compressed through the auto-encoder, high-quality results can be produced in a small capacity with only the main characteristics of the data (eg, characteristic variables).

2 is a diagram illustrating an example of building an autoencoder model of a new unfair claim detection apparatus according to an embodiment of the present application.

Referring to FIG. 2, the new pattern detection unit 130 may detect a new pattern of unfair claims from error values of each of the normal autoencoder model and the abnormal autoencoder model. The new pattern detection unit 130 may calculate a normal error value from the normal autoencoder model for learning the normal feature variable, and may calculate the abnormal error value from the abnormal autoencoder move for learning the abnormal feature variable, respectively. The new pattern detection unit 130 may detect a new pattern of unfair claims by using the normal error value and the abnormal error value.

3 is a diagram showing a comparison of a normal error value and an abnormal error value of a new unfair claim detection device according to an embodiment of the present application, and FIG. 4 is an example of labeling of a new unfair claim detection device according to an embodiment of the present application It is a diagram showing.

The new pattern detection unit 130 may utilize the difference between the normal error value and the abnormal error value to detect a new pattern of unfair claims. Referring to FIG. 3, the new pattern detection unit 130 is the difference between the error value (NAE Error) of the normal auto encoder and the error value (FAE Error) of the Fraud Auto Encoder (NAE − FAE) can be calculated. The new pattern detection unit 130 may set a critical section for detecting a new pattern prior to detecting the new pattern. Referring to FIG. 4, the critical section may include a normal, a suspected group, and an abnormality, and the new pattern detection unit 130 is a threshold for classifying each of a normal, a suspected group, and an abnormality based on the difference (NAE-FAE). The interval, that is, the range of the threshold value can be set. The threshold section may be set by a user, and a threshold section previously set for each difference (NAE-FAE) may be used. As another example, the threshold section using the difference (NAE-FAE) may be set based on a threshold value calculated as (|NAE-FAE|) 0.1.

Referring to FIG. 4, the new pattern detection unit 130 may classify and label a critical section based on a difference between an error value of a normal autoencoder model and an error value of an abnormal autoencoder model. In addition, the new pattern detection unit 130 may detect a new pattern from learning data (or feature variable) associated with an error value labeled as a suspect group. Exemplarily, when the error value labeled as a suspect group exceeds the threshold value ((|NAE-FAE|) 0.1), the new pattern detection unit 130 generates a new pattern based on the feature variable associated with the error value. Can be defined.

5 is a diagram illustrating an example of detection of a new unfair pattern by a new unfair claim detection apparatus according to an embodiment of the present application.

According to the exemplary embodiment of the present application, the new pattern detection unit 130 may classify a new pattern of unfair claims based on a clustering algorithm based on unsupervised learning that inputs a feature variable. Specifically, unsupervised learning refers to an algorithm that learns while analyzing or clustering the data itself rather than constructing data for learning. Since this is a known matter, a detailed description will be omitted.

The new pattern detection unit 130 may cluster characteristic variables into a plurality of claim patterns based on the frequency of the insurance claim details data. As an example, the new pattern detection unit 130 may classify a new pattern of unfair claims based on at least one of a K-means clustering algorithm and a Self-Organizing-Maps (SOM) algorithm and EM & Canopy algorithm. The K-means clustering algorithm is a traditional classification technique that iteratively subdivides the target group into K clusters based on the mean value (similarity) of the distance, and the SOM algorithm learns the input pattern of the training set with weights based on artificial neural networks. It is a technique of clustering. In addition, the EM & Canopy algorithm refers to a method of clustering by updating parameter values through an iterative process starting from the one with the maximum possibility with a given initial value.

The new pattern detection unit 130 may detect a new pattern through the clustering algorithm based on the frequency of the feature variable. FIG. 2 shows patterns clustered according to the frequency of the feature variables, and when the frequencies of feature variables belonging to different patterns are similar, as shown in FIG. 2, they may be classified into the same or similar colors (blue). That is, when the colors between patterns are the same or similar, it can be said that the degree of separation between clusters is low. In addition, it can be determined that the claim pattern clustered at a similar frequency among characteristic variables is a clustered claim pattern, that is, a normal claim pattern rather than an unfair claim based on the frequency of the characteristic variable generated by normal insurance claim.

Meanwhile, in the case of a pattern clustered by feature variables having a similar frequency and feature variables having a different frequency, as shown in FIG. 5, a different color (red) from the aforementioned claim pattern may be used. Since this claim pattern is different from the frequency of the characteristic variable of the normal claim pattern (for example, when the number of FP changes is relatively larger than that of the normal claim pattern), it may be a new pattern of unfair claims, and the new pattern. It can be detected by the detection unit 130.

The new unfair claim determination unit 140 may determine an unfair claim of a new insurance claim based on an unfair claim detection model for learning a new pattern. Specifically, the new unfair claim determination unit 140 may determine an unfair claim of a new insurance claim based on a new pattern learning model and a new pattern based on a supervised learning-based classification algorithm. The classification algorithm includes a decision algorithm, and the decision algorithm may be, for example, a decision tree algorithm, but is not limited thereto. The new unfair claim determination unit 140 constructs a supervised learning-based unfair claim detection model that inputs a new pattern of unfair claims, a pattern of mood, and a characteristic variable based on the insurance claim details data of the new insurance claim and is not labeled. It is possible to determine whether a new insurance claim is unfairly claimed.

6 is a diagram illustrating a flow of a method for detecting a new unfair claim according to an embodiment of the present application.

The method for detecting a new unfair claim shown in FIG. 6 may be performed by the new unfair claim detection apparatus 100 described with reference to FIGS. 1 to 5 above. Therefore, even if omitted below, the description of the new unfair claim detection apparatus 100 through FIGS. 1 to 5 may be equally applied to FIG. 6.

Referring to FIG. 6, in step S610, the data preprocessing unit 110 may pre-process and format insurance claim details data including unstructured data, and may extract characteristic variables from the standardized insurance claim details data. For example, in the case of customer data, the standardization of insurance claim data refers to quantifying the income level of the customer, the number of hospitals visited by the customer, and the number of disease reasons the customer has applied for.

In step S620, the model building unit 120 builds a normal autoencoder model that learns the normal characteristic variable derived from the normal insurance claim details data and the abnormal autoencoder model that learns the abnormal characteristic variable derived from the abnormal insurance claim details data. I can. According to an embodiment of the present application, the data preprocessing unit 110 may distinguish between normal and abnormal insurance claim details data in the process of calculating characteristic variables through pre-processing of insurance claim details data that are not classified into normal and abnormal. I can. Specifically, the data preprocessing unit 110 standardizes each element (claim data, contract data, etc.) included in the above-described insurance claim detail data, and makes a corresponding insurance claim according to whether or not a preset threshold value is exceeded for each of the elements. It is possible to distinguish whether the historical data is normal insurance claim history data or abnormal insurance claim history data. In addition, the model building unit 120 may learn characteristic variables derived from the normal insurance claim details data and the abnormal insurance claim details data separated by the data preprocessing unit 110. The model building unit 120 may build a normal autoencoder model and an abnormal autoencoder model based on an unsupervised learning-based autoencoder algorithm.

In step S630, the new pattern detection unit 130 may detect a new pattern of unfair claims from error values of each of the normal autoencoder model and the abnormal autoencoder model. The new pattern detection unit 130 may calculate a normal error value from the normal autoencoder model for learning the normal feature variable, and may calculate the abnormal error value from the abnormal autoencoder move for learning the abnormal feature variable, respectively. The new pattern detection unit 130 may detect a new pattern of unfair claims by using the normal error value and the abnormal error value. The new pattern detection unit 130 may utilize the difference between the normal error value and the abnormal error value to detect a new pattern of unfair claims. Referring to FIG. 3, the new pattern detection unit 130 is the difference between the error value (NAE Error) of the normal auto encoder and the error value (FAE Error) of the Fraud Auto Encoder (NAE − FAE) can be calculated.

The new pattern detection unit 130 may set a critical section for detecting a new pattern prior to detecting the new pattern. The critical section may include a normal, a suspect group, and an abnormality, and the new pattern detection unit 130 divides each of a normal, a suspect group, and an abnormality based on the difference (NAE-FAE), that is, a range of a threshold value. Can be set. The threshold section may be set by a user, and a threshold section previously set for each difference (NAE-FAE) may be used. As another example, the threshold section using the difference (NAE-FAE) may be set based on a threshold value calculated as (|NAE-FAE|) 0.1.

The new pattern detection unit 130 may classify and label a critical section based on a difference between an error value of a normal autoencoder model and an error value of an abnormal autoencoder model. In addition, the new pattern detection unit 130 may detect a new pattern from learning data (or feature variable) associated with an error value labeled as a suspect group. Exemplarily, when the error value labeled as a suspect group exceeds the threshold value ((|NAE-FAE|) 0.1), the new pattern detection unit 130 generates a new pattern based on the feature variable associated with the error value. Can be defined.

In step S640, the new unfair claim determination unit 140 may determine the unfair claim of the new insurance claim based on the unfair claim detection model for learning a new pattern. Specifically, the new unfair claim determination unit 140 may determine an unfair claim of a new insurance claim based on a new pattern learning model and a new pattern based on a supervised learning-based classification algorithm. The classification algorithm includes a decision algorithm, and the decision algorithm may be, for example, a decision tree algorithm, but is not limited thereto. The new unfair claim discrimination unit 140 constructs a supervised learning-based unfair claim detection model that inputs a new pattern of unfair claim, a pattern that has been changed, and characteristic variables based on the insurance claim details data of the new insurance claim. It is possible to determine whether the billing details are unfairly billed.

According to an embodiment of the present disclosure, a method for detecting a new unfair claim may be implemented in the form of a program command that can be executed through various computer means and recorded in a computer-readable medium. The computer-readable medium may include program instructions, data files, data structures, and the like alone or in combination. The program instructions recorded on the medium may be specially designed and configured for the present invention, or may be known and usable to those skilled in computer software. Examples of computer-readable recording media include magnetic media such as hard disks, floppy disks, and magnetic tapes, optical media such as CD-ROMs and DVDs, and magnetic media such as floptical disks. -A hardware device specially configured to store and execute program instructions such as magneto-optical media, and ROM, RAM, flash memory, and the like. Examples of program instructions include not only machine language codes such as those produced by a compiler but also high-level language codes that can be executed by a computer using an interpreter or the like. The above-described hardware device may be configured to operate as one or more software modules to perform the operation of the present invention, and vice versa.

The foregoing description of the present application is for illustrative purposes only, and those of ordinary skill in the art to which the present application pertains will be able to understand that it is possible to easily transform it into other specific forms without changing the technical spirit or essential features of the present application. Therefore, it should be understood that the embodiments described above are illustrative and non-limiting in all respects. For example, each component described as a single type may be implemented in a distributed manner, and similarly, components described as being distributed may also be implemented in a combined form.

The scope of the present application is indicated by the claims to be described later rather than the detailed description, and all changes or modified forms derived from the meaning and scope of the claims and their equivalent concepts should be interpreted as being included in the scope of the present application.

100: new unfair claim detection device
110: data preprocessor
120: model building section
130: new pattern detection unit
140: New unfair claim determination unit
150: database

Claims (15)

In the new unfair claim detection device,
A data preprocessing unit that pre-processes the insurance claim details data to derive and refine characteristic variables required for artificial intelligence learning;
A model building unit for constructing a normal autoencoder model for learning normal characteristic variables derived from normal insurance claim details data and an abnormal autoencoder model for learning abnormal characteristic variables derived from abnormal insurance claim details data;
A new pattern detection unit detecting a new pattern of unfair claims from an error value of each of the normal autoencoder model and the abnormal autoencoder model; And
Including a new unfair claim determination unit for determining the unfair claim of the new insurance claim based on the unfair claim detection model for learning the new pattern,
The new pattern detection unit,
A critical section for detecting a new pattern is set, wherein the critical section includes a normal, a suspect group, and an abnormality,
Based on the difference between the error value of the normal auto-encoder model and the error value of the abnormal auto-encoder model, set a range of threshold values that divide the critical section,
Labeling is performed in the critical section corresponding to the difference between the error value of the normal autoencoder model and the error value of the abnormal autoencoder model,
If there is an error value exceeding the threshold value among the respective error values labeled as the suspicious group, the new pattern is detected from a feature variable associated with the error value exceeding the threshold value,
The new unfair claim determination unit,
The new pattern learning model based on the supervised learning-based classification algorithm and to determine the unfair claim of a new insurance claim based on the new pattern. The method of claim 1,
Further comprising a database for storing insurance claim history data including unstructured data,
The insurance claim detail data includes at least one of claim data, contract data, payment data, insurance agent data, and customer data. The method of claim 1,
The model building unit,
To construct the normal autoencoder model and the abnormal autoencoder model based on an unsupervised learning-based autoencoder algorithm. delete delete delete delete In the method for detecting a new unfair claim by a new unfair claim detection device,
Pre-processing the insurance claim details data, deriving and refining characteristic variables required for artificial intelligence learning;
Constructing a normal autoencoder model for learning a normal characteristic variable derived from the normal insurance claim details data and an abnormal autoencoder model for learning the abnormal characteristic variable derived from the abnormal insurance claim details data;
Detecting a new pattern of unfair claims from error values of each of the normal autoencoder model and the abnormal autoencoder model; And
Including the step of determining an unfair claim of a new insurance claim based on the unfair claim detection model for learning the new pattern,
The step of detecting the new pattern,
A critical section for detecting a new pattern is set, wherein the critical section includes a normal, a suspect group and an abnormality,
Based on the difference between the error value of the normal auto-encoder model and the error value of the abnormal auto-encoder model, set a range of threshold values that divide the critical section
Labeling is performed in the critical section corresponding to the difference between the error value of the normal autoencoder model and the error value of the abnormal autoencoder model,
If there is an error value exceeding the threshold value among the respective error values labeled as the suspicious group, the new pattern is detected from a feature variable associated with the error value exceeding the threshold value,
The step of determining the unfair claim of the new insurance claim,
A new pattern learning model based on a supervised learning-based classification algorithm, and to determine an unfair claim of a new insurance claim based on the new pattern. The method of claim 8,
Further comprising the step of storing the insurance claim details data including unstructured data,
The insurance claim details data includes at least one of claim data, contract data, payment data, insurance agent data, and customer data. The method of claim 8,
The step of building the model,
To construct the normal autoencoder model and the abnormal autoencoder model based on an unsupervised learning-based autoencoder algorithm, a new unfair claim detection method. delete delete delete delete A computer-readable recording medium storing a program for executing the method of claim 8 on a computer.

Images