KR20220083197A - Blockchain based Insurance fraud detection system through sharing information of claims for insurance - Google Patents

Abstract

The present invention relates to an insurance fraud detection system that shares insurance claim information and calculates the possibility of insurance fraud by comparing similarities. A blockchain-based insurance claim that encrypts and shares insurance customer's insurance claim information among multiple insurance companies in a convergence network, extracts insurance claim patterns from shared information, and calculates the possibility of insurance fraud by comparing the patterns It relates to an insurance fraud detection system through information sharing.

Description

Blockchain based Insurance fraud detection system through sharing information of claims for insurance}

The present invention relates to an insurance fraud detection system that shares insurance claim information and calculates the possibility of insurance fraud by comparing similarities. More specifically, the Blockchain-based Internet of Things (IoT) is integrated A blockchain-based insurance claim that encrypts and shares insurance customer's insurance claim information among multiple insurance companies in a convergence network, extracts insurance claim patterns from shared information, and calculates the possibility of insurance fraud by comparing the patterns It relates to an insurance fraud detection system through information sharing.

Insurance fraud includes intentionally causing an accident or fabricating a non-insured accident as an insurance accident, fraudulently overcharging insurance money to receive a large amount of insurance money due to an insurance accident, and There are acts such as changing the driver or vehicle and making a claim for insurance.

Insurance claims through such insurance fraud are becoming more intelligent day by day, and as the number of insurance claims continues to increase, insurance money is being paid unfairly.

To prevent this, insurance companies are developing methods such as accumulating data related to insurance fraud cases and extracting insurance fraud patterns through big data. The method of extracting insurance fraud patterns through machine learning using big data is a method of predicting whether there is a possibility of fraud by extracting certain patterns seen in insurance fraud. In order to use this method, insurance fraud patterns must be exposed in the existing big data. However, new types of insurance fraud that are not disclosed in big data are not recognized as insurance fraud patterns. There is a problem that cannot but suffer. In addition, it is difficult to predict that it is an insurance fraud even in the case of an insurance claim that is intentional but does not have a fraud pattern by the general public who has no history of insurance fraud in the past.

In addition, most insurance frauds do not stop at once, but change the method and make a common claim from multiple companies. Methods for determining whether insurance fraud is detected by extracting patterns of incidents are being studied. There are difficulties.

In addition, as personal information about a specific user is stored on a third party server, there is a risk of personal information being exposed to a third party, and as the server information accumulates, it can become a target for hackers to leak personal information from the server. Additional problems may arise.

The present invention has been devised to solve the above problems, and an object of the present invention is to share insurance claim information of an insurance company between insurance companies in real time, and to use machine learning for information that has a possibility of fraud among the shared insurance claim information. It is to provide an insurance fraud detection system by sharing insurance claim information based on block chain that calculates and predicts insurance fraud immediately.

In particular, insurance fraud detection through blockchain-based insurance claim information sharing, which protects insurance customers' personal information by encrypting insurance claim information and sharing it through block chain, and at the same time solves the problem of personal information leakage due to server information accumulation to provide a system.

The insurance fraud detection system through the sharing of insurance claim information based on the block chain according to an embodiment of the present invention is a block chain node that stores and manages insurance company information and insurance claim information in the block chain form in the block chain-based platform. ; and an insurance fraud detection application that is registered in the blockchain network and determines insurance claim patterns and insurance fraud potential; including, wherein the blockchain node issues a predetermined common seed key in common to a plurality of insurance company applications, and the insurance fraud detection application includes information shared by the plurality of insurance company applications through the blockchain node. It is characterized by judging the possibility of insurance fraud through a similarity analysis between the two.

In addition, the information shared by the plurality of insurance company applications through the block chain node includes a hash code generated through a hash function that is a one-way function based on the common seed key.

In addition, the blockchain node requests authentication whether the insurance company is a normal insurance company from the insurance company authentication server registered in the blockchain network based on the insurance company information received from the insurance company application, and the insurance company When the insurance company is authenticated as a normal insurance company from the authentication server, the insurance company is registered as a client and the common seed key is issued.

In addition, it is characterized in that the blockchain node obtains insurance claim information from the insurance company application, and notifies the insurance fraud detection application of the occurrence of an insurance claim event.

In addition, the insurance claim information includes personal information and accident details or damage details in the form of keywords. It is characterized in that it is encrypted so that there is no

In addition, it is characterized in that the personal information-related information and the personal information among the keywords of the accident details or damage details are encrypted with a hash code using the hash function.

In addition, the personal information is characterized in that it includes at least one of an insurance party, an applicant, a beneficiary, and an insurance subject.

In addition, the accident details or damage history is characterized in that it includes at least one information of the occurrence time, the occurrence place, and the type of insurance accident.

In addition, when the insurance fraud detection application is notified of the occurrence of the insurance claim event, it extracts an insurance claim pattern from the insurance claim information through machine learning, and shares the extracted insurance claim pattern with the block chain node in the prior art The insurance fraud possibility is calculated based on the insurance claim patterns, and the extracted insurance claim pattern and the insurance fraud potential are shared through the block chain node.

In addition, the insurance claim pattern is characterized in that it includes information based on the personal information encrypted with the hash code.

In addition, the insurance fraud detection application is characterized in that the higher the degree of similarity between the extracted insurance claim pattern and each of the previously shared insurance claim patterns, the higher the likelihood of insurance fraud is determined.

In addition, the insurance fraud detection application calculates the degree of similarity between the extracted insurance claim pattern and each of the previously shared insurance claim patterns, and when the number of similarities greater than or equal to a predetermined criterion is greater than or equal to the predetermined reference number, the possibility of insurance fraud is increased. It is characterized by being judged to be high.

In addition, when comparing insurance claim patterns to calculate the similarity, the similarity weight of information based on the personal information in the comparison target insurance claim patterns is higher than the similarity weight of information other than the personal information.

In addition, when the blockchain node obtains insurance fraud case information from the insurance company application, it notifies the insurance fraud detection application of the occurrence of an insurance fraud case event, and when obtaining insurance claim information from the insurance company application, the It is characterized by notifying the occurrence of an insurance claim event to the insurance fraud detection application.

In addition, when the insurance fraud detection application receives notification of the occurrence of the insurance fraud case event, the insurance fraud case information is machine-learned to extract the insurance fraud pattern, shared through the block chain node, and the insurance claim event Upon receiving the notification, the insurance claim information is machine-learned to extract the insurance claim pattern and shared through the block chain node.

In addition, the insurance fraud detection application calculates the possibility of insurance fraud based on the extracted insurance claim pattern, the conventional insurance fraud pattern shared through the block chain node, and the insurance claim patterns, and the extracted insurance money It is characterized in that the claim pattern and the possibility of insurance fraud are shared through the blockchain node.

In addition, the insurance fraud detection application calculates a first degree of similarity between the extracted insurance money claim pattern and the shared insurance fraud pattern, and if the first degree of similarity is higher than a predetermined criterion, it is determined that the probability of insurance fraud is high; When the first degree of similarity is less than or equal to a predetermined standard, a second degree of similarity between the extracted insurance claim pattern and each of the previously shared insurance claim patterns is calculated, and when the number of second similarities equal to or greater than the predetermined standard is greater than or equal to the predetermined reference number, insurance It is characterized in that the probability of fraud is high.

The insurance fraud detection system through the sharing of insurance claim information based on the block chain of the present invention according to the above configuration extracts a newly emerging insurance fraud pattern by sharing insurance accident claim related information, examining similarity, and extracting a certain pattern has a useful effect.

In addition, since insurance accident claim information is stored using hash code conversion, there is an effect of preventing leakage of personal information due to sharing of claim information.

In addition, the block chain in which insurance accident claim information is shared has the effect of blocking access to uncertified insurance companies or hackers by allowing only certified insurance companies to use it as it is linked with the authentication server.

In addition, since insurance claim information encrypted through hashing cannot be decrypted, it can be used usefully only by insurance companies that want to check whether insurance fraud has occurred. there is

1 is a block diagram of a system for detecting insurance fraud through sharing insurance claim information based on a block chain according to an embodiment of the present invention;
2 is a flowchart of a detection method using an insurance fraud detection system through sharing insurance claim information based on a block chain according to a first embodiment of the present invention;
3 is a flowchart of a detection method using an insurance fraud detection system through sharing insurance claim information based on a block chain according to a second embodiment of the present invention;

Hereinafter, an embodiment of the present invention as described above will be described in detail with reference to the drawings.

1 is a block diagram of an insurance fraud detection system 1000 (hereinafter, 'insurance fraud detection system') through block chain-based insurance claim information sharing according to an embodiment of the present invention. As shown, the insurance fraud detection system 1000 includes an insurance company application 100 , a block chain node 200 , an authentication server 300 , and an insurance fraud detection application 400 .

The insurance company application 100 may be provided singly or in plurality to correspond to the insurance company(s). The insurance company application 100 is a client connected to the block chain node 200, and enters insurance claim information into the block chain node 200 to make the insurance claim information unique with other insurance companies, and to request information on the possibility of insurance fraud do. The insurance company application 100 may receive an account for accessing the block chain node 200 and a seed key for sharing only necessary information (insurance claim information) from the block chain node 200 . To this end, the insurance company application 100 includes application logic for authentication.

The blockchain node 200 is configured to store and manage insurance company information and insurance claim information in a blockchain format on a blockchain-based platform. In particular, the block chain node 200 may be applied with smart contract technology through Ethereum. As an example, the block chain node 200 stores insurance claim information input through the insurance company application 100 as a client connected to the Ethereum network, and stores insurance fraud possibility information extracted from the insurance claim information. It is provided to the company application (100).

The authentication server 300 is a server for authenticating the insurance company application 100 that wants to access the block chain node 200, and is configured to block the access of insurance companies or hackers except for designated insurance companies. The authentication server 300 checks the certificate provided by the insurance company application 100 from the blockchain node 200 , and registers the result (with or without authentication) in the blockchain node 200 . Accordingly, the authentication server 300 includes authentication server logic that performs authentication.

The insurance fraud detection application 400 extracts an insurance claim pattern through insurance claim information registered in the block chain network and stored in the block chain node 200, and compares it with a conventional insurance claim pattern to determine the possibility of insurance fraud. is composed To this end, the insurance fraud detection application 400 includes a machine learning block. The insurance fraud detection application 400 has an account for inquiring insurance claim information stored in the block chain node 200 and registering the extracted insurance claim pattern in the block chain node 200 . The machine learning block compares and analyzes the correlation and similarity between the extracted insurance claim patterns to determine the possibility of insurance fraud. The machine learning block may be embedded in the smart contract of the block chain node 200, and more preferably, a method of interworking from the outside for smart contract performance and independence of machine learning pattern development language may be applied.

Hereinafter, various embodiments of the insurance fraud detection method using the insurance fraud detection system 1000 as described above will be described in detail with reference to the drawings.

- Example 1

2 is a flowchart of a detection method using an insurance fraud detection system through sharing insurance claim information based on a block chain according to a first embodiment of the present invention.

As shown, the insurance fraud detection method largely includes a company registration step (S10), an insurance claim information registration step (S30), and an insurance fraud determination step (S50).

The company registration step (S10) is a step in which various insurance companies register accounts in the blockchain node 200 to share insurance claim information with each other.

More specifically, the insurance company that wants to be authenticated performs the step S11 of requesting to register its own account to the blockchain node 200 through the insurance company application 100 . The blockchain node 200 performs a step (S12) of requesting the authentication server 300 to authenticate the insurance company account. The authentication server 300 performs a step (S13) of comparing the DB information with the insurance company based on the stored insurance company DB. When the insurance company is authenticated as a normal insurance company through the DB information, the authentication server 300 performs a step S14 of approving the authentication request of the blockchain node 200 . When the authentication request is approved through the authentication server 300, the blockchain node 200 registers the insurance company as a client, and performs a step (S15) of issuing an account and a common seed key to the insurance company. This is to prepare for the case where a general fake user, not an insurance company, attempts to register in the blockchain network by impersonating an insurance company. When the blockchain node 200 is linked with the authentication server 300 , it can be linked with an external server through the Oracleize method, which is one of the data transmission means connecting the blockchain and external API.

The insurance claim information registration step (S30) is configured to register and share insurance claim information with the blockchain node 200 when the customer's insurance claim details are generated.

More specifically, the insurance company application 100 outputs at least some of the insurance claim information as a one-way hash function based on a common seed key, and generates a hash code through the function (S31). do. The hash code (Y) may be defined as follows.

Y= Hash(Info, SeedKey)

(here, Y: hash code, Hash: hash function, Info: claim information, SeedKey: seed key)

In addition, a plurality of hash codes (Y) may be generated to correspond to a plurality of pieces of information included in the insurance claim information, and Yn hash codes may be generated from Y1 according to the number of information (Info) included in the insurance claim information. can That is, as shown below, any one insurance claim information P1 may include hash codes Y11, Y12 ... Y1n as many as the number of information included.

P1∋ {Y11, Y12, … , Y1n}

In this case, the insurance claim information includes customer information such as the insurance party, applicant, insurance beneficiary, insurance subject, and accident-related person, which is the customer's personal information. In addition, the accident details or damage details may further include information such as the occurrence time, place, and accident type. As described above, the customer's personal information and keywords that can leak personal information are stored in an encrypted hash code based on the insurance company's seed key. When a hash code including insurance claim information is generated, the hash code is uploaded to the block chain node 200 and shared with a plurality of insurance companies that have an account of the block chain node 200 (S32) do. Insurance claim information in the hash code can only be shared between insurance companies that have been issued a common seed key. In addition, since the hash code, which is Encrypted Customer Information (ECI), is shared in the blockchain node 200, insurance companies that share insurance claim information create encrypted customer information (ECI) in the same way to commit insurance fraud. The possibility can be checked, but the encrypted customer information (ECI) cannot be decrypted, so the personal information contained in the hash code cannot be extracted backwards. In other words, only the party with the same insurance claim information can check whether insurance fraud is possible, and even certified insurance companies cannot verify personal information unless they encrypt and register the same insurance claim information as in the case. Therefore, even if the insurance claim details are shared among insurance companies, there is no risk of personal information leakage, and only the relevant insurance company can verify the contents, and even the related insurance company cannot decrypt the encrypted customer information (ECI), so insurance fraud is possible. It is only used to determine whether or not there is a problem, and fundamentally blocks the possibility of leakage of customer personal information.

When the hash code is registered, the blockchain node 200 notifies the occurrence of an insurance claim event to the insurance fraud detection application 400 to perform the insurance fraud determination step (S50) through the insurance fraud detection application 400 The occurrence notification step (S33) is performed.

The insurance fraud determination step (S50) is to extract an insurance claim pattern from insurance claim information registered in the blockchain node 200 through the insurance fraud detection application 400, predict insurance fraud through the extracted information, and register the related information. is composed To this end, the insurance fraud detection application 400 performs a step (S51) of extracting an insurance claim pattern from the insurance claim information through machine learning. At this time, the extracted insurance claim pattern includes information based on personal information encrypted with a hash code, and is characterized in that it cannot be decrypted. Calculating the degree of similarity between the extracted insurance money claim pattern and the conventional insurance claim pattern (S52) is performed. In addition, a step (S53) of calculating the possibility of insurance fraud based on the calculated similarity is performed. Here, in the process of extracting the insurance claim pattern, the identity of the party, the applicant, the beneficiary, etc., the correlation with the insurance vehicle, the correlation with the place of occurrence, etc. are checked.

Also, by extracting patterns for accident techniques and comparing them with existing shared patterns, the possibility is examined. For example, it can be determined that the higher the degree of similarity between the extracted insurance claim pattern and each of the previously shared insurance claim patterns is, the higher the likelihood of insurance fraud.

As an example, any one insurance claim information (P1) is defined as follows,

P1= (Y11, Y12, Y13, …,Y1n)

After defining the other insurance claim information (P2) as follows,

P2= (Y21, Y22, Y23, …,Y2n)

Y11, Y12, Y13, ... of P1. , Y21, Y22, Y23, … of P2 corresponding to Y1n , Y2n is compared with each other, and as the number of identical hash codes increases, it is determined that P1 and P2 are more likely to be fraudulent information, and Y11, Y12, Y13, ... ,Y1n and Y21, Y22, Y23, … ,Y2n can be extracted as a fraud pattern.

As another example, the degree of similarity between the extracted insurance claim pattern and each of the previously shared insurance claim patterns is calculated, and when the number of similarities greater than or equal to a predetermined threshold is greater than or equal to the predetermined reference number, it can be determined that the likelihood of insurance fraud is high. . Also, when comparing insurance claim patterns to calculate similarity, a similarity weight of information based on the personal information in the comparison target insurance claim patterns may be higher than a similarity weight of information other than the personal information. That is, when the similarity of information based on personal information and the similarity of information other than personal information are the same, it can be determined that the former has a high possibility of insurance fraud. Finally, the insurance company performs the step (S54) of registering the insurance claim pattern and the calculated insurance fraud possibility in the block chain node 200 to be used for calculating the insurance fraud possibility. Furthermore, the insurance fraud detection application 400 is a machine learning module in the field of artificial intelligence that is registered as a client module externally to improve performance problems, and sharing of insurance claim information occurs through event processing of the block chain node 200. Each time, the results are quickly compared and registered in the blockchain node 200.

- Example 2

3 is a flowchart of a detection method using an insurance fraud detection system through sharing insurance claim information based on a block chain according to a second embodiment of the present invention.

As shown, the insurance fraud detection method largely includes a company registration step (S10), an insurance fraud case registration step (S20), an insurance claim information registration step (S30), and an insurance fraud determination step (S60).

Since the company registration step (S10) is the same as the company registration step (S10) of the first embodiment described above, a detailed description thereof will be omitted.

The insurance fraud case registration step (S20) is configured to register and share insurance fraud case information held by insurance companies in the blockchain node 200.

More specifically, the insurance company application 100 outputs at least some of the insurance fraud case information as a hash function that is a one-way function based on a common seed key, and generates a first hash code including the function. Step S21 is performed. In this case, the insurance fraud case information includes customer personal information, insurance beneficiaries, and customer information related to accidents, and the details of the insurance accident and the insurance accident are registered as keywords. The customer's personal information and keywords that can leak personal information are stored in the first encrypted hash code of the insurance company as described above. Step ( S22) is performed. Insurance fraud case information in the first hash code can be shared only between insurance companies that have been issued a common seed key. In addition, since the first hash code, which is encrypted customer information (ECI), is shared in the block chain node 200, insurance companies that share insurance fraud case information create encrypted customer information (ECI) in the same way to prevent insurance fraud. Although it is possible to check whether or not the encrypted customer information (ECI) cannot be decrypted, the personal information included in the first hash code cannot be extracted backwards. In other words, only the party with similar insurance claim information can check whether insurance fraud is possible through the insurance fraud case information, and even certified insurance companies cannot verify personal information unless the insurance claim information is encrypted and registered. do. When the first hash code is registered, the block chain node 200 notifies the occurrence of an insurance fraud case event to the insurance fraud detection application 400 and immediately extracts the insurance fraud pattern through the insurance fraud detection application 400 ( S24) is performed.

The insurance claim information registration step ( S30 ) is the same as the insurance claim information registration step ( S30 ) of the above-described first embodiment, and thus a detailed description thereof will be omitted.

In the insurance fraud determination step (S60), an accident pattern of insurance claim information registered in the blockchain node 200 is extracted through the insurance fraud detection application 400, and the extracted information is compared with an insurance fraud pattern or a conventional insurance claim pattern. It is configured to register related information by predicting insurance fraud through To this end, the insurance fraud detection application 400 performs a step (S51) of extracting an insurance claim pattern from the insurance claim information through machine learning. Next, a step (S62) of calculating a first degree of similarity is performed by comparing the extracted insurance claim pattern with the extracted insurance fraud pattern. In addition, a step (S63) of calculating a second degree of similarity is performed by comparing the extracted insurance claim pattern with the conventional insurance claim pattern. In addition, a step (S64) of calculating the possibility of insurance fraud based on the calculated first and second similarities is performed. For example, if the first degree of similarity is higher than a predetermined standard, it is determined that the probability of insurance fraud is high. The degree of similarity is calculated, and when the number of times having the second degree of similarity equal to or greater than a predetermined criterion is equal to or greater than the predetermined reference number, it may be determined that the probability of insurance fraud is high. Here, in the process of extracting the insurance claim pattern, the identity of the party, the applicant, the beneficiary, etc., the correlation with the insurance vehicle, the correlation with the place of occurrence, etc. are checked. Also, by extracting patterns for accident techniques and comparing them with existing shared patterns, the possibility is examined. Finally, insurance companies share the insurance fraud possibility by registering the insurance claim pattern and the calculated insurance fraud possibility in the blockchain node 200 (S65). Furthermore, the insurance fraud detection application 400 is a machine learning module in the field of artificial intelligence that is registered as a client module externally to improve performance problems, and sharing of insurance claim information occurs through event processing of the block chain node 200. Each time, the results are quickly compared and registered in the blockchain node 200.

The technical idea should not be construed as being limited to the above-described embodiment of the present invention. Various modifications can be made at the level of those skilled in the art without departing from the gist of the present invention as claimed in the claims. Accordingly, such improvements and modifications fall within the protection scope of the present invention as long as it is apparent to those skilled in the art.

100: insurance company application(s)
200 : Blockchain Node
300: authentication server
400: insurance fraud detection application

Claims (17)

In the insurance fraud detection system through the sharing of insurance claim information based on block chain,
a block chain node that stores and manages insurance company information and insurance claim information in a block chain form on the block chain-based platform; and
Insurance fraud detection application that is registered in the blockchain network and determines insurance claim patterns and insurance fraud potential; including,
The blockchain node issues a predetermined common seed key in common to a plurality of insurance company applications,
The insurance fraud detection application determines the possibility of insurance fraud through similarity analysis between information shared by the plurality of insurance company applications through the block chain node.
An insurance fraud detection system through the sharing of insurance claim information based on blockchain.
The method of claim 1,
The information shared by the plurality of insurance company applications through the block chain node includes a hash code generated through a hash function that is a one-way function based on the common seed key.
An insurance fraud detection system through the sharing of insurance claim information based on blockchain.
3. The method of claim 2,
The blockchain node is
Based on the information of the insurance company received from the insurance company application, the insurance company authentication server registered in the blockchain network requests authentication whether the insurance company is a normal insurance company,
When the insurance company is authenticated as a normal insurance company from the insurance company authentication server, registering the insurance company as a client and issuing the common seed key
An insurance fraud detection system through the sharing of insurance claim information based on blockchain.
3. The method of claim 2,
The blockchain node,
Obtain insurance claim information from the insurance company application,
Notifying the insurance fraud detection application of the occurrence of an insurance claim event
An insurance fraud detection system through the sharing of insurance claim information based on blockchain.
5. The method of claim 4,
The insurance claim information is
Including accident details or damage history in the form of personal information and keywords,
Among the keywords of the accident details or damage details, personal information-related information and the personal information are encrypted so that the insurance fraud detection application cannot decipher it.
An insurance fraud detection system through the sharing of insurance claim information based on blockchain.
6. The method of claim 5,
Among the keywords of the accident details or damage details, personal information-related information and the personal information are encrypted with a hash code using the hash function
An insurance fraud detection system through the sharing of insurance claim information based on blockchain.
6. The method of claim 5,
The personal information includes at least one of the insured party, the applicant, the beneficiary, and the subject of insurance
An insurance fraud detection system through the sharing of insurance claim information based on blockchain.
6. The method of claim 5,
The accident details or damage details include information on at least one of the time of occurrence, the place of occurrence, and the type of insurance accident
An insurance fraud detection system through the sharing of insurance claim information based on blockchain.
5. The method of claim 4,
The insurance fraud detection application,
Upon notification of the occurrence of the insurance claim event,
extracting the insurance claim pattern from the insurance claim information through machine learning,
Calculate the possibility of insurance fraud based on the extracted insurance claim pattern and the previously shared insurance claim patterns through the block chain node,
Sharing the extracted insurance claim pattern and the possibility of insurance fraud through the blockchain node
An insurance fraud detection system through the sharing of insurance claim information based on blockchain.
10. The method of claim 9,
The insurance claim pattern includes information based on personal information encrypted with the hash code
An insurance fraud detection system through the sharing of insurance claim information based on blockchain.
10. The method of claim 9,
The insurance fraud detection application,
Determining that the higher the degree of similarity between the extracted insurance claim pattern and each of the previously shared insurance claim patterns, the higher the likelihood of insurance fraud
An insurance fraud detection system through the sharing of insurance claim information based on blockchain.
12. The method of claim 11,
The insurance fraud detection application,
Calculating the degree of similarity between the extracted insurance claim pattern and each of the previously shared insurance claim patterns, and determining that insurance fraud is highly probable when the number of similarities greater than or equal to a predetermined standard is greater than or equal to the predetermined reference number
An insurance fraud detection system through the sharing of insurance claim information based on blockchain.
12. The method of claim 11,
When comparing insurance claim patterns to calculate the similarity, the similarity weight of information based on the personal information in the comparison target insurance claim patterns is higher than the similarity weight of information other than the personal information
An insurance fraud detection system through the sharing of insurance claim information based on blockchain.
3. The method of claim 2,
The blockchain node,
When insurance fraud case information is obtained from the insurance company application, notifying the insurance fraud detection application of the occurrence of an insurance fraud case event;
Notifying the occurrence of an insurance claim event to the insurance fraud detection application when obtaining insurance claim information from the insurance company application
An insurance fraud detection system through the sharing of insurance claim information based on blockchain.
15. The method of claim 14,
The insurance fraud detection application,
Upon receiving notification of the occurrence of the insurance fraud case event, the insurance fraud case information is machine-learned to extract insurance fraud patterns and shared through the blockchain node;
Upon receiving the notification of the insurance claim event, machine learning the insurance claim information to extract the insurance claim pattern and share it through the blockchain node
An insurance fraud detection system through the sharing of insurance claim information based on blockchain.
16. The method of claim 15,
The insurance fraud detection application,
Calculating the possibility of insurance fraud based on the extracted insurance claim pattern, the conventional insurance fraud pattern shared through the block chain node, and the insurance claim patterns,
Sharing the extracted insurance claim pattern and the possibility of insurance fraud through the blockchain node
An insurance fraud detection system through the sharing of insurance claim information based on blockchain.
17. The method of claim 16,
The insurance fraud detection application,
A first degree of similarity is calculated between the extracted insurance claim pattern and the shared insurance fraud pattern, and when the first degree of similarity is higher than a predetermined criterion, it is determined that the possibility of insurance fraud is high;
When the first similarity is less than or equal to a predetermined criterion, a second degree of similarity between the extracted insurance claim pattern and each of the previously shared insurance claim patterns is calculated, and when the number of second similarities equal to or greater than the predetermined criterion is greater than or equal to the predetermined reference number , which is judged to be highly likely to be fraudulent
An insurance fraud detection system through the sharing of insurance claim information based on blockchain.

Images