Method for automatically extracting core content of insurance responsibility

Abstract

The invention provides a method for automatically extracting core content of insurance responsibility. The method comprises the following steps: extracting character information in an insurance clause file; 2, retrieving a guarantee responsibility content paragraph from the extracted character information; a third step of matching the standard responsibility field by using cosine similarity for the guarantee responsibility content paragraph so as to obtain matched guarantee responsibility content; 4, reordering the matched guarantee responsibility content by using a Page Rank algorithm to obtain first ordered content; 5, using the standard responsibility field data to form training data; sixthly, key information is automatically extracted from the first sorted content through training data to obtain scattered key information; and a seventh step of integrating the obtained scattered key information so as to obtain core content for guaranteeing responsibility.

Description

Technical Field

[0001] This invention relates to the field of insurance, and specifically to a method for automatically extracting the core content of insurance coverage. Background Technology

[0002] Insurance liability refers to the items for which an insurance company assumes responsibility for compensation or payment of insurance benefits. Specifically, insurance liability refers to the scope of risk that the insurer assumes as stipulated in the insurance contract, and the liability for compensation in the event of an insured event. This includes damages, liability compensation, insurance benefits, rescue expenses, relief expenses, and litigation costs. After the insured signs the insurance contract and pays the premium, the scope of liability stipulated in the insurance contract becomes the insurer's liability. In the event of property damage or personal injury within the scope of insurance liability, the insurer is responsible for compensation or payment of insurance benefits. The scope of the insurer's liability for compensation or payment of insurance benefits includes: the damage occurring within the scope of insurance liability; the insurance liability occurring within the insurance period; and the limit being the insured amount. Therefore, insurance liability is both the insurer's responsibility to provide protection and the basis and scope for compensation and payment of insurance benefits; it is also the insured's responsibility to claim protection and the basis and scope for obtaining compensation or benefits. Different types of insurance have different insurance liabilities, which are mainly divided into basic liabilities, special liabilities, and exclusions.

[0003] It is evident that the core content of insurance coverage is an extremely important aspect of the insurance contract. However, the core content of insurance coverage varies among individual insurance contracts, and both insurance professionals and policyholders desire to quickly and accurately understand it; however, this is not easy.

[0004] The goal is to provide a technical solution that can quickly and accurately extract the core content of liability protection. Summary of the Invention

[0005] The technical problem to be solved by this invention is to address the aforementioned deficiencies in the existing technology by providing an automatic method for extracting the core content of insurance coverage. This method employs methods such as cosine similarity matching, PageRank algorithm sorting, and automatic extraction of key information from training data, and applies these methods to the field of insurance clauses to form a fast and effective automatic method for extracting the core content of insurance coverage.

[0006] According to the present invention, a method for automatically extracting the core content of insurance coverage is provided, comprising:

[0007] First step: Extract textual information from the insurance policy documents;

[0008] The second step: Retrieve the paragraphs containing the guarantee responsibility from the extracted text information;

[0009] The third step: For the paragraphs containing the guarantee responsibility content, use cosine similarity to match the standard responsibility fields to obtain the matched guarantee responsibility content;

[0010] Fourth step: Use the PageRank algorithm to re-rank the matched guarantee liability content to obtain the first ranked content;

[0011] Step 5: Use standard responsibility field data to generate training data;

[0012] Step 6: Use the training data to automatically extract key information from the first sorted content to obtain scattered key information;

[0013] Step 7: Integrate the scattered key information obtained to obtain the core content of the guarantee responsibility.

[0014] Preferably, the method for automatically extracting the core content of insurance coverage according to the present invention further includes:

[0015] Step 8: Use the HITS algorithm to reorder the matched guarantee liability content to obtain the second sorted content;

[0016] Step 9: Automatically extract key information from the second sorted content using training data to obtain auxiliary scattered key information;

[0017] Step 10: Adjust the core content of the obtained guarantee liability using supplementary, scattered key information.

[0018] Preferably, the paragraphs containing the content on guarantee responsibility are retrieved from the extracted text information using the words "guarantee responsibility" or similar words.

[0019] Preferably, the standard liability field data includes standard terminology used for coverage in the insurance contract.

[0020] Preferably, the matching algorithm used to match the standard responsibility field is the cosine similarity algorithm. Attached Figure Description

[0021] A more complete understanding of the invention and its accompanying advantages and features will be more readily apparent from the accompanying drawings and the following detailed description, wherein:

[0022] Figure 1 A flowchart illustrating an automatic extraction method for the core content of insurance coverage according to a preferred embodiment of the present invention is shown.

[0023] It should be noted that the accompanying drawings are for illustrative purposes only and are not intended to limit the invention. Note that the drawings illustrating structures may not be drawn to scale. Furthermore, in the drawings, identical or similar elements are labeled with the same or similar reference numerals. Detailed Implementation

[0024] To make the content of this invention clearer and easier to understand, the content of this invention will be described in detail below with reference to specific embodiments and accompanying drawings.

[0025] Figure 1 A flowchart illustrating an automatic extraction method for the core content of insurance coverage according to a preferred embodiment of the present invention is shown.

[0026] like Figure 1 As shown, the method for automatically extracting the core content of insurance coverage according to a preferred embodiment of the present invention includes:

[0027] Step S1: Extract textual information from the insurance policy documents;

[0028] Step S2: Retrieve the paragraph containing the guarantee responsibility from the extracted text information; for example, using "guarantee responsibility" or similar words, retrieve the paragraph containing the guarantee responsibility from the extracted text information.

[0029] Step S3: For the paragraph containing the guarantee liability content, use cosine similarity to match the standard liability field to obtain the matched guarantee liability content;

[0030] Step S4: Use the PageRank algorithm to re-sort the matched guarantee liability content to obtain the first sorted content;

[0031] Step S5: Use standard liability field data to generate training data; for example, standard liability field data includes standard terminology for coverage liability in insurance contracts.

[0032] Step S6: Automatically extract key information from the first sorted content using the training data to obtain scattered key information;

[0033] Step S7: Integrate the scattered key information obtained to obtain the core content of the guarantee responsibility.

[0034] Preferably, such as Figure 1 As shown, the method for automatically extracting the core content of insurance coverage according to a preferred embodiment of the present invention further includes:

[0035] Step S8: Use the HITS algorithm to reorder the matched guarantee liability content to obtain the second sorted content;

[0036] Step S9: Automatically extract key information from the second sorted content using the training data to obtain auxiliary scattered key information;

[0037] Step 10: Adjust the core content of the obtained guarantee liability using the scattered key information provided as an auxiliary tool.

[0038] <Specific examples of extracting text information>

[0039] For example, insurance terms are originally recorded in a PDF file, which needs to be parsed into a TXT text file.

[0040] This example uses a lazy parsing strategy, parsing only when needed to reduce time and memory usage. The following classes are primarily required:

[0041] PDFParser: Extracts data from a file;

[0042] PDFDocument: Used to store data;

[0043] PDFPageInterpreter: Processes page content;

[0044] PDFDevice: Converts it to a predefined format;

[0045] PDFResourceManager: Used to store shared content.

[0046] <Specific examples of retrieving coverage information>

[0047] Therefore, the PDF file is parsed to obtain a TXT text file, and the guarantee responsibility title is used to search, match, and extract the content from the TXT text file.

[0048] <Specific example of matching standard responsibility fields>

[0049] Each coverage module contains one or more coverage benefits, such as minor illness insurance benefits, moderate illness insurance benefits, critical illness insurance benefits, extended malignant tumor insurance benefits, extended insurance benefits for specific cardiovascular and cerebrovascular diseases, death or total disability insurance benefits, etc. These coverage benefit names need to be categorized and standardized into uniform fields to prevent situations where the same benefit has different names.

[0050] For example, the matching algorithm used is the cosine similarity algorithm. Cosine distance, also known as cosine similarity, uses the cosine of the angle between two vectors a and b in vector space as a measure of the difference between the two individuals. The closer the cosine value is to 1, the more similar the two vectors are, which is called "cosine similarity". The closer the cosine value is to 0, the less similar the two vectors are, which means the two strings are less similar.

[0051] in,

[0052] Taking critical illness insurance benefits as an example, the results shown in the table below can be obtained:

[0053] Standard fields Existing fields in the terms Similarity Critical Illness Insurance Benefit Critical Illness Insurance Benefit 1.00 Critical Illness Insurance Benefit First Critical Illness Insurance Benefit 0.78 Critical Illness Insurance Benefit Critical Illness Insurance Benefit Payment 0.78 Critical Illness Insurance Benefit Ordinary critical illness insurance benefits 0.78 Critical Illness Insurance Benefit Critical Illness Insurance Coverage 0.75 Critical Illness Insurance Benefit First Critical Illness Insurance Benefit 0.70

[0054] For example, a threshold of 0.70 can be set, meaning that when the similarity is greater than or equal to 0.70, the two fields to be compared are considered similar and can be used as a unified field; otherwise, they are not similar.

[0055] <Specific examples of reordering the content of protection responsibilities>

[0056] The previous step used cosine similarity to obtain the names of similar guarantee liability fields. Then, the complete guarantee liability content corresponding to the field can be obtained, and the PageRank algorithm can be used to reorder the statements of the content.

[0057] For example, the text before sorting is:

[0058] "If the insured is diagnosed for the first time with one or more of the 100 critical illnesses defined in this contract due to accidental injury or, after the waiting period, due to reasons other than accidental injury, and is initially diagnosed by a specialist doctor at a medical institution recognized by the insurance company, the insurance company will pay 100% of the basic sum insured under this contract as critical illness insurance benefit. If you did not choose to purchase the optional insurance liability 'Extended Insurance Benefit for Malignant Tumors' or 'Extended Insurance Benefit for Specific Cardiovascular and Cerebrovascular Diseases' at the time of application, this insurance liability will terminate. If you chose to purchase the optional insurance liability 'Extended Insurance Benefit for Malignant Tumors' or 'Extended Insurance Benefit for Specific Cardiovascular and Cerebrovascular Diseases' at the time of application, this insurance liability will terminate. After the insurance company pays the critical illness insurance benefit, you will not need to continue to pay the premiums payable according to the contract from the date of the insured's initial diagnosis of a critical illness. The cash value of this contract will be reduced to zero, and the insurance company will no longer bear any other insurance liabilities except for the 'Extended Insurance Benefit for Malignant Tumors' or 'Extended Insurance Benefit for Specific Cardiovascular and Cerebrovascular Diseases' under this contract. The cumulative number of critical illness insurance benefits under this contract is limited to two."

[0059] After sorting, we get:

[0060] "If the insured is diagnosed for the first time with one or more of the 100 critical illnesses defined in this contract due to accidental injury or, after the waiting period, due to reasons other than accidental injury, and is initially diagnosed by a specialist doctor at a medical institution recognized by the insurance company, the insurance company will pay 100% of the basic sum insured under this contract as critical illness insurance benefits. The cumulative number of critical illness insurance benefits under this contract is limited to two. If you did not choose to purchase the optional insurance liability 'Extended Insurance Benefit for Malignant Tumors' or 'Extended Insurance Benefit for Specific Cardiovascular and Cerebrovascular Diseases' at the time of application, this insurance liability will terminate. If you chose to purchase the optional insurance liability 'Extended Insurance Benefit for Malignant Tumors' or 'Extended Insurance Benefit for Specific Cardiovascular and Cerebrovascular Diseases' at the time of application, this insurance liability will terminate. After the insurance company pays the critical illness insurance benefits, you will no longer need to pay the premiums payable according to the contract from the date of the insured's initial diagnosis of a critical illness. The cash value of this contract will be reduced to zero, and the insurance company will no longer bear any other insurance liabilities except for the 'Extended Insurance Benefit for Malignant Tumors' or 'Extended Insurance Benefit for Specific Cardiovascular and Cerebrovascular Diseases' under this contract."

[0061]

[0062] Taking the coverage of critical illness insurance as an example, the key information is shown in red in the image above. Key information data for various types of insurance clauses can be manually created (human-defined) as training data, and then automatically extracted using this training data. An example of the result after automatically extracting key information from the coverage details shown in the image above is shown below:

[0063] "100 Major Diseases"

[0064] The critical illness insurance benefit will be paid out at 100% of the basic sum insured under the contract.

[0065] The maximum number of times the critical illness insurance benefit can be paid under this contract is two.

[0066] <Specific examples of integrating key information to form the core content of safeguard responsibility>

[0067] Integrate the scattered key information obtained in the previous step to form the core content of the coverage. Taking critical illness insurance and general medical insurance as examples, the core content of the coverage is: "100 critical illnesses, not grouped. Two payouts, 100% of the basic sum insured" and "Sum insured of 1 million, deductible of 10,000. With social security settlement: 100% reimbursement ratio. Without social security settlement: 70% reimbursement ratio."

[0068] It should be noted that, unless otherwise specified, the terms "first," "second," "third," etc., in the specification are used only to distinguish the various components, elements, and steps in the specification, and are not used to indicate the logical or sequential relationships between the various components, elements, and steps.

[0069] It is understood that although the present invention has been disclosed above with reference to preferred embodiments, these embodiments are not intended to limit the present invention. For any person skilled in the art, many possible variations and modifications can be made to the technical solutions of the present invention based on the disclosed technical content, or equivalent embodiments can be modified accordingly, without departing from the scope of the present invention. Therefore, any simple modifications, equivalent changes, and modifications made to the above embodiments based on the technical essence of the present invention without departing from the content of the present invention shall still fall within the protection scope of the present invention.

Claims

1. A method for automatically extracting the core content of insurance coverage, characterized in that... include: First step: Extract textual information from the insurance policy documents; The second step: Retrieve the paragraphs containing the guarantee responsibility from the extracted text information; The third step: For the paragraphs containing the guarantee responsibility content, use cosine similarity to match the standard responsibility fields to obtain the matched guarantee responsibility content; Fourth step: Use the PageRank algorithm to re-rank the matched guarantee liability content to obtain the first ranked content; Step 5: Use standard responsibility field data to generate training data; Step 6: Use the training data to automatically extract key information from the first sorted content to obtain scattered key information; Step 7: Integrate the scattered key information obtained to obtain the core content of the guarantee responsibility.

2. The method for automatically extracting the core content of insurance coverage as described in claim 1, characterized in that... Also includes: Step 8: Use the HITS algorithm to reorder the matched guarantee liability content to obtain the second sorted content; Step 9: Automatically extract key information from the second sorted content using training data to obtain auxiliary scattered key information; Step 10: Adjust the core content of the obtained guarantee liability using supplementary, scattered key information.

3. The method for automatically extracting the core content of insurance coverage as described in claim 1 or 2, characterized in that, Using phrases like "guarantee responsibility" or similar words, retrieve paragraphs containing the guarantee responsibility content from the extracted text information.

4. The method for automatically extracting the core content of insurance coverage as described in claim 1 or 2, characterized in that, The standard liability field data includes the standard terminology used for coverage in insurance contracts.

5. The method for automatically extracting the core content of insurance coverage as described in claim 1 or 2, characterized in that, The matching algorithm used to match the standard responsibility fields is the cosine similarity algorithm.

Images