CN-121998780-A - Automatic insurance claim settlement processing method and system based on artificial intelligence

CN121998780ACN 121998780 ACN121998780 ACN 121998780ACN-121998780-A

Abstract

The invention relates to the technical field of insurance, and particularly discloses an automatic insurance claim settlement processing method and system based on artificial intelligence. According to the method, the claim settlement request data are automatically screened through artificial intelligence and are divided into structured data (insurance numbers, amounts and the like) and unstructured data (photos, texts and the like), basic insurance difference values of insurance information and historical records are calculated based on the structured data, accident scene flow nodes are extracted from the unstructured data, flow node difference values are obtained through analysis and processing step normalization, text difference values are obtained through comparison of traffic and alarm records and claim texts, the three comprehensive difference values are synthesized to calculate comprehensive abnormal risk difference values, if the three comprehensive abnormal risk difference values exceed a threshold value, manual auditing is carried out, otherwise, automatic claim settlement is carried out, and the method achieves multidimensional risk quantification, and achieves automatic efficiency and accurate identification of cheating insurance risks.

Inventors

ZHENG JUNLIN
XU DONGJUN
WU MINGZHI
LI JINGLUN

Assignees

浙江优财云链科技有限公司

Dates

Publication Date: 20260508
Application Date: 20260320
Priority Date: 20260106

Claims (10)

1. An artificial intelligence-based insurance claim automation processing method is characterized by comprising the following steps: Acquiring insurance claim request data of an applicant, and reporting and screening the insurance claim request data based on artificial intelligence to obtain screening insurance claim request data, wherein the screening insurance claim request data comprises structured data and unstructured data; Acquiring application basic data and historical claim settlement record data according to the structured data, and acquiring basic application difference values according to the application basic data and the historical claim settlement record data; Acquiring accident scene data and text description data according to the unstructured data, acquiring field processing flow node data according to the accident scene data, and acquiring a difference value of each processing flow node according to the field processing flow node data; Acquiring traffic record data and report record data according to the text description data, and acquiring text difference values according to the traffic record data and the report record data; Calculating a comprehensive abnormal risk difference value according to the basic application difference value, the plurality of processing flow node difference values and the text difference value; Judging whether the comprehensive abnormal risk difference value is larger than a preset threshold value or not; if the comprehensive abnormal risk difference value is larger than a preset threshold value, judging that the applicant has a cheating protection risk, generating a manual checking instruction, and performing manual checking on the insurance claim settlement request data according to the manual checking instruction; If the comprehensive abnormal risk difference value is not greater than a preset threshold value, acquiring corresponding claim settlement amount according to the application basic data, and automatically carrying out insurance claim settlement processing on the applicant according to the claim settlement amount.
2. The automated processing method for insurance claims based on artificial intelligence according to claim 1, wherein the step of screening the insurance claim request data based on artificial intelligence to obtain screened insurance claim request data includes: Performing data standardization preprocessing on the insurance claim request data to obtain initial standardization data, wherein the data standardization preprocessing comprises missing value filling and abnormal value removing; inputting the initial normalized data into a pre-trained multi-mode data classification model to obtain a data completeness score and a data type label, wherein the data type label comprises a structured data label and an unstructured data label; Judging whether the data completeness score is lower than a preset completeness threshold value or not; If the data completeness score is lower than a preset completeness threshold, generating a data complement instruction and sending the data complement instruction to the applicant terminal; If the data completeness score is not lower than a preset completeness threshold, dividing the initial normalized data into structured data and unstructured data according to the data type label, and combining the structured data and the unstructured data together to form screening insurance claim settlement request data.
3. The automated artificial intelligence based insurance claim processing method of claim 1, wherein said step of obtaining a base insurance difference value from said application base data and said historical claim record data includes: Acquiring key insurance parameters according to the insurance basic data, wherein the key insurance parameters comprise insurance policy effective time, dangerous seed type, insurance amount and insured person age; Cosine similarity matching is carried out on the effective time length of the policy, the type of the dangerous seed, the amount of applied insurance and the age of the insured person according to the historical claim settlement record data, so that a plurality of historical related matching claim settlement record data are obtained; Extracting the first K principal components from the plurality of history related matching claim settlement record data to obtain a plurality of principal component claim settlement data, wherein the principal component claim settlement data comprises a first history claim settlement times, a first average claim settlement amount and a first latest claim settlement time interval; obtaining a second historical claim count, a second average claim amount and a second latest claim time interval of the applicant according to the application basic data, and obtaining basic application difference values according to the second historical claim count, the second average claim amount, the second latest claim time interval, a plurality of first historical claim count, a plurality of first average claim amount and a plurality of first latest claim time intervals.
4. The automated processing method of claim 1, wherein the step of obtaining field process node data from the incident scene data and obtaining each process node difference value from the field process node data comprises: Performing image recognition and time sequence analysis on the accident scene data to obtain a key processing flow node and a scene processing step sequence, wherein the key processing flow node comprises an alarm time point, a traffic police arrival time point, a scene investigation completion time point and a vehicle dragging time point; Standard accident handling data are acquired, and standard time intervals and processing step constraint sequences between each two key processing flow nodes are acquired according to the standard accident handling data; Sequentially obtaining a plurality of time intervals between each other according to the alarm time point, the traffic police arrival time point, the on-site investigation completion time point and the vehicle towing-off time point, wherein the time intervals comprise an alarm-traffic police arrival time interval, a traffic police arrival-on-site investigation completion time interval and an on-site investigation completion-vehicle towing-off time interval; Comparing the standard time intervals with the alarm-traffic police arrival time interval, the traffic police arrival-on-site investigation completion time interval and the on-site investigation completion-vehicle towing-off time interval in sequence to obtain a plurality of time node difference values; And obtaining a step deviation coefficient according to the field processing step sequence and the processing step constraint sequence, and obtaining the node difference value of each processing flow according to the step deviation coefficient and the time node difference values.
5. The automated artificial intelligence based insurance claim processing method according to claim 1, wherein the step of obtaining text variance values from the traffic record data and the newspaper record data includes: Text analysis and key information extraction are carried out on the traffic record data to obtain a first accident description element set, wherein the first accident description element set comprises a first accident time, a first accident place, a first vehicle identifier and a first responsibility identification result; Text analysis and key information extraction are carried out on the report record data to obtain a second accident description element set, wherein the second accident description element set comprises a second accident time, a second accident place, a second vehicle identifier and a second responsibility identification result; performing element alignment matching on the first accident description element set and the second accident description element set to obtain a plurality of matching element pairs; Obtaining a plurality of local difference degrees according to the matching element pairs, and obtaining an average local difference degree according to the local difference degrees; Calculating a standard local difference according to the local differences and the average local difference, calculating a local difference coefficient according to the standard local difference and the average local difference, and taking the local difference coefficient as a text difference value.
6. The automated artificial intelligence based insurance claim 1, wherein said step of calculating a composite anomaly risk difference value from said base applied difference value, a plurality of said process flow node difference values, and said text difference value, includes: Normalizing the basic application difference value, the plurality of processing flow node difference values and the text difference value to obtain a basic application normalized difference value, a plurality of processing flow node normalized difference values and the text normalized difference value; weighting and calculating the basic insuring normalized difference value, the normalized difference values of a plurality of processing flow nodes and the normalized difference value of the text to obtain a basic abnormal risk difference value; acquiring a risk adjustment factor of a current claim settlement business scene; And acquiring an adjusted abnormal risk difference value according to the risk adjustment factor and the basic abnormal risk difference value, and taking the adjusted abnormal risk difference value as a comprehensive abnormal risk difference value.
7. An artificial intelligence based insurance claim automation processing system, comprising: the system comprises a first acquisition module, a second acquisition module and a third acquisition module, wherein the first acquisition module is used for acquiring insurance claim settlement request data of an applicant, and carrying out declaration screening on the insurance claim settlement request data based on artificial intelligence to obtain screening insurance claim settlement request data, wherein the screening insurance claim settlement request data comprises structured data and unstructured data; The second acquisition module is used for acquiring the application basic data and the historical claim settlement record data according to the structured data, and acquiring basic application difference values according to the application basic data and the historical claim settlement record data; the third acquisition module is used for acquiring accident scene data and text description data according to the unstructured data, acquiring field processing flow node data according to the accident scene data and acquiring a difference value of each processing flow node according to the field processing flow node data; The fourth acquisition module is used for acquiring traffic record data and report record data according to the text description data and acquiring text difference values according to the traffic record data and the report record data; the calculation module is used for calculating a comprehensive abnormal risk difference value according to the basic application difference value, the plurality of processing flow node difference values and the text difference value; The judging module is used for judging whether the comprehensive abnormal risk difference value is larger than a preset threshold value or not; if the comprehensive abnormal risk difference value is larger than a preset threshold value, judging that the applicant has a cheating protection risk, generating a manual checking instruction, and performing manual checking on the insurance claim settlement request data according to the manual checking instruction; If the comprehensive abnormal risk difference value is not greater than a preset threshold value, acquiring corresponding claim settlement amount according to the application basic data, and automatically carrying out insurance claim settlement processing on the applicant according to the claim settlement amount.
8. An artificial intelligence based insurance claim automation processing system according to claim 7, the method is characterized in that the first acquisition module comprises the following steps: The first acquisition unit is used for carrying out data standardization preprocessing on the insurance claim settlement request data to obtain initial standardization data, wherein the data standardization preprocessing comprises missing value filling and abnormal value removing; The training unit is used for inputting the initial standardized data into a pre-trained multi-mode data classification model to obtain a data completeness score and a data type label, wherein the data type label comprises a structured data label and an unstructured data label; the judging unit is used for judging whether the data completeness score is lower than a preset completeness threshold value or not; If the data completeness score is lower than a preset completeness threshold, generating a data complement instruction and sending the data complement instruction to the applicant terminal; If the data completeness score is not lower than a preset completeness threshold, dividing the initial normalized data into structured data and unstructured data according to the data type label, and combining the structured data and the unstructured data together to form screening insurance claim settlement request data.
9. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor implements the steps of the method of any of claims 1 to 6 when the computer program is executed.
10. A computer readable storage medium, on which a computer program is stored, characterized in that the computer program, when being executed by a processor, implements the steps of the method according to any one of claims 1 to 6.

Description

Automatic insurance claim settlement processing method and system based on artificial intelligence Technical Field The invention relates to the technical field of insurance, in particular to an automatic insurance claim settlement processing method and system based on artificial intelligence. Background The insurance claim is taken as a key link of the insurance service, and directly related to the operation efficiency, risk control capability and customer satisfaction of an insurance company, the traditional insurance claim settlement flow is highly dependent on manual processing, and a claim settlement technician needs to manually review various materials submitted by an applicant; Secondly, with the rapid growth of insurance business, insurance fraud is accompanied in the process of insurance claim settlement, however, the existing identification of the risk of fraud protection mainly depends on the experience of individuals of claim settlement specialists, and manual audit is unavoidable, and for elaborate fraudulent insurance actions, especially fraudulent cases of false information manufactured in a plurality of links by utilizing information asymmetry, systematic and quantitative detection means are lacked, so that the problem of higher rate of missed detection of the risk of fraud protection is caused, and for this reason, an automatic processing method of insurance claim settlement based on artificial intelligence is required to solve the problem. Disclosure of Invention The invention aims to provide an artificial intelligence-based insurance claim automatic processing method and system, which are used for solving the technical problems in the background technology. In order to achieve the purpose, the invention provides the following technical scheme that the automatic insurance claim settlement processing method based on artificial intelligence comprises the following steps: Acquiring insurance claim request data of an applicant, and reporting and screening the insurance claim request data based on artificial intelligence to obtain screening insurance claim request data, wherein the screening insurance claim request data comprises structured data and unstructured data; Acquiring application basic data and historical claim settlement record data according to the structured data, and acquiring basic application difference values according to the application basic data and the historical claim settlement record data; Acquiring accident scene data and text description data according to the unstructured data, acquiring field processing flow node data according to the accident scene data, and acquiring a difference value of each processing flow node according to the field processing flow node data; Acquiring traffic record data and report record data according to the text description data, and acquiring text difference values according to the traffic record data and the report record data; Calculating a comprehensive abnormal risk difference value according to the basic application difference value, the plurality of processing flow node difference values and the text difference value; Judging whether the comprehensive abnormal risk difference value is larger than a preset threshold value or not; if the comprehensive abnormal risk difference value is larger than a preset threshold value, judging that the applicant has a cheating protection risk, generating a manual checking instruction, and performing manual checking on the insurance claim settlement request data according to the manual checking instruction; If the comprehensive abnormal risk difference value is not greater than a preset threshold value, acquiring corresponding claim settlement amount according to the application basic data, and automatically carrying out insurance claim settlement processing on the applicant according to the claim settlement amount. Preferably, the step of reporting and screening the insurance claim request data based on artificial intelligence to obtain screened insurance claim request data includes: Performing data standardization preprocessing on the insurance claim request data to obtain initial standardization data, wherein the data standardization preprocessing comprises missing value filling and abnormal value removing; inputting the initial normalized data into a pre-trained multi-mode data classification model to obtain a data completeness score and a data type label, wherein the data type label comprises a structured data label and an unstructured data label; Judging whether the data completeness score is lower than a preset completeness threshold value or not; If the data completeness score is lower than a preset completeness threshold, generating a data complement instruction and sending the data complement instruction to the applicant terminal; If the data completeness score is not lower than a preset completeness threshold, dividing the initial normalized data into structured data and unstructured data ac