Case fraud identification method and system, computer equipment and storage medium

By obtaining and converting user voices, using the fraud identification model to automatically judge the fraud risk of auto insurance cases, the efficiency and experience problems of insurance companies when handling auto insurance fraud cases are solved, and claims efficiency and user satisfaction are improved.

CN119088968APending Publication Date: 2024-12-06CHINA PING AN PROPERTY INSURANCE CO LTD
View PDF 2 Cites 0 Cited by

Patent Information

Application Number
CN202411210646.2
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-08-30
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

Insurers face problems such as low claims efficiency, difficulty in identifying fraud, poor user experience and high human resources costs when dealing with auto insurance fraud cases. They need a method to automatically identify fraud risks to improve claims efficiency and improve user experience.

Method used

By obtaining the user's voice, converting it into text using NLP technology, and using the preset fraud identification model to determine whether the case has a risk of fraud. If it exists, it will be marked as an abnormal case and sent to the claim agent for processing, including manual Review and on-site inspection.

Benefits of technology

It realizes the automatic identification of fraud risks in user cases, improves claims efficiency and user experience, reduces human resources investment, and significantly improves the automation level and overall efficiency of claims processing.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119088968A_ABST
    Figure CN119088968A_ABST
Patent Text Reader

Abstract

The invention discloses a case fraud identification method and system, computer equipment and a storage medium, and relates to the technical field of information processing, and the method comprises the steps: obtaining the call voice of a target user about a target case; converting the call voice into a corresponding target text; and based on the target text, through a preset fraud identification model, determining whether the target case has a fraud risk. Through the method, whether the user case has the fraud risk can be automatically identified, the claim settlement efficiency is improved, and the user experience is improved.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the field of information processing technology, and in particular to a case fraud identification method, system, computer equipment and non-volatile computer-readable storage medium. Background Art

[0002] At present, with the rapid development of the insurance industry, especially the auto insurance market, the types of insurance products are becoming increasingly rich and the user base is expanding. However, with this comes the frequent occurrence of auto insurance fraud cases, which not only bring huge economic losses to insurance companies, but also affect the claims experience of normal users.

[0003] Currently, insurance companies rely on claims agents to manually review and process cases reported by users. However, the number of claims agents is limited, and facing the growing number of cases, there are the following problems:

[0004] 1. Claims settlement efficiency: When handling a large number of user cases, claims agents find it difficult to complete the review and claims settlement of all cases in a short period of time due to human resource constraints, resulting in low processing efficiency.

[0005] 2. Fraud identification problem: Auto insurance fraud cases are often hidden and complex. It is difficult for claims agents to accurately identify all fraud cases within a limited time, which increases the risk of insurance companies.

[0006] 3. User experience issues: After reporting a case, users often need to wait for a long time for the insurance company to handle the case. This long wait seriously affects the user's service experience;

[0007] 4. Human resource costs: As the number of cases increases, insurance companies need to invest more claims seats to handle cases, which undoubtedly increases the company's human resource costs;

[0008] Based on this, how to provide a case fraud identification method, system, computer equipment and non-volatile computer-readable storage medium that can automatically identify whether a user's case has a fraud risk, improve claims efficiency, and improve user experience is a problem that urgently needs to be solved by technical personnel in this field. Summary of the invention

[0009] In view of the above-mentioned shortcomings of the prior art, the purpose of the present invention is to provide a case fraud identification method, system, computer equipment and non-volatile computer-readable storage medium that can be used in financial technology or other related fields, aiming to solve the problem of how to automatically identify whether a user's case has a fraud risk, improve claims efficiency, and improve user experience.

[0010] In order to achieve the above object, the present invention adopts the following technical solutions:

[0011] A case fraud identification method, comprising:

[0012] Obtain the target user's call audio regarding the target case;

[0013] Converting the call speech into corresponding target text;

[0014] Based on the target text, a preset fraud identification model is used to determine whether the target case has a fraud risk.

[0015] In a further technical solution, the case fraud identification method, wherein the step of obtaining the target user's call voice regarding the target case, includes:

[0016] Processing incoming calls from target users regarding reporting target cases through a telephone platform, and transferring the target users to corresponding AI customer service;

[0017] Relevant call audio is obtained through the call process between the AI ​​customer service and the target user regarding the target case.

[0018] In a further technical solution, the case fraud identification method, wherein the converting of the call speech into the corresponding target text comprises:

[0019] Through NLP speech conversion technology, the call speech is converted into the target text corresponding to the target case.

[0020] In a further technical solution, the case fraud identification method, wherein the step of judging whether the target case has a fraud risk based on the target text by using a preset fraud identification model, includes:

[0021] Pre-built fraud detection models;

[0022] Inputting the target text into the fraud identification model to generate a fraud identification result of the target case;

[0023] Based on the fraud identification result, determine whether there is a fraud risk in the target case.

[0024] In a further technical solution, the case fraud identification method, wherein the pre-built fraud identification model comprises:

[0025] Collect some user case information;

[0026] After data cleaning and feature selection of the user case information, a corresponding target data set is generated;

[0027] After training and verifying the preset machine learning model using the target data set, a corresponding fraud identification model is generated.

[0028] In a further technical solution, the case fraud identification method, wherein judging whether the target case has a fraud risk based on the fraud identification result, further includes:

[0029] If there is a risk of fraud in the target case, the target case will be sent to the corresponding claims seat for processing as an abnormal case.

[0030] In a further technical solution, the case fraud identification method, wherein if the target case has a fraud risk, the target case is sent as an abnormal case to a corresponding claims seat for processing, further comprising:

[0031] In response to the broadcast instruction of the call voice by the claims agent, the claims agent manually reviews the target case;

[0032] When the review result shows that the target case is indeed an abnormal case, an investigation instruction is sent to the corresponding investigator, so that the investigator conducts an on-site investigation of the target case.

[0033] A case fraud identification system, comprising:

[0034] An acquisition module is used to acquire the target user's call voice regarding the target case;

[0035] A conversion module, used for converting the call speech into a corresponding target text;

[0036] The judgment module is used to judge whether there is a fraud risk in the target case based on the target text through a preset fraud recognition model.

[0037] A computer device, wherein the computer device comprises at least one processor; and

[0038] a memory communicatively connected to the at least one processor; wherein,

[0039] The memory stores a computer program that can be executed by the at least one processor. When the computer program is executed by the at least one processor, the case fraud identification method as described in any one of the above items can be implemented.

[0040] A non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by at least one processor, it can implement the case fraud identification method as described in any one of the above items.

[0041] Compared with the prior art, the present invention provides a case fraud identification method, system, computer device and non-volatile computer-readable storage medium, wherein the method includes: obtaining the target user's call voice about the target case; converting the call voice into the corresponding target text; based on the target text, judging whether the target case has a fraud risk through a preset fraud identification model. In this way, the method of the present invention can automatically identify whether the user's case has a fraud risk, improve claims efficiency, and improve user experience. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0043] Figure 1 A flowchart of a case fraud identification method provided in an embodiment of the present invention.

[0044] Figure 2 A schematic diagram of the functional modules of a case fraud identification system provided in an embodiment of the present invention.

[0045] Figure 3 A schematic diagram of the hardware structure of a computer device provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0046] In order to make the purpose, technical solution and effect of the present invention clearer and more specific, the present invention is further described in detail with reference to the accompanying drawings and examples. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not used to limit the present invention.

[0047] In the description of the present invention, the terms "include", "including", "have", "contain", etc. are all open terms, which mean including but not limited to. The descriptions with reference to the terms "one embodiment", "a specific embodiment", "some embodiments", "for example", etc. mean that the specific features, structures or characteristics described in conjunction with the embodiment or example are included in at least one embodiment or example of the present application. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures or characteristics described may be combined in a suitable manner in any one or more embodiments or examples. The order of steps involved in each embodiment is used to schematically illustrate the implementation of the present application, and the order of steps therein is not limited and can be appropriately adjusted as needed.

[0048] Various non-limiting embodiments of the present invention are described in detail below in conjunction with the accompanying drawings.

[0049] At present, with the rapid development of the insurance industry, especially the auto insurance market, the types of insurance products are becoming increasingly rich and the user base is expanding. However, with this comes the frequent occurrence of auto insurance fraud cases, which not only bring huge economic losses to insurance companies, but also affect the claims experience of normal users.

[0050] Currently, insurance companies rely on claims agents to manually review and process cases reported by users. However, the number of claims agents is limited, and facing the growing number of cases, there are the following problems:

[0051] 1. Claims settlement efficiency: When handling a large number of user cases, claims agents find it difficult to complete the review and claims settlement of all cases in a short period of time due to human resource constraints, resulting in low processing efficiency.

[0052] 2. Fraud identification problem: Auto insurance fraud cases are often hidden and complex. It is difficult for claims agents to accurately identify all fraud cases within a limited time, which increases the risk of insurance companies.

[0053] 3. User experience issues: After reporting a case, users often need to wait for a long time for the insurance company to handle the case. This long wait seriously affects the user's service experience;

[0054] 4. Human resource costs: As the number of cases increases, insurance companies need to invest more claims seats to handle cases, which undoubtedly increases the company's human resource costs;

[0055] Based on this, how to provide a case fraud identification method, system, computer equipment and non-volatile computer-readable storage medium that can automatically identify whether a user's case has a fraud risk, improve claims efficiency, and improve user experience is a problem that urgently needs to be solved by technical personnel in this field.

[0056] To solve the above problems, see Figure 1 The embodiment of the present invention provides a case fraud identification method, wherein the method comprises the steps of:

[0057] S100, obtaining a target user's call voice regarding a target case;

[0058] S200, converting the call speech into a corresponding target text;

[0059] S300: Based on the target text, determine whether the target case has a fraud risk by using a preset fraud identification model.

[0060] Furthermore, in the case fraud identification method, the step S100 of obtaining the target user's call voice regarding the target case includes:

[0061] Processing incoming calls from target users regarding reporting target cases through a telephone platform, and transferring the target users to corresponding AI customer service;

[0062] Relevant call audio is obtained through the call process between the AI ​​customer service and the target user regarding the target case.

[0063] In the specific implementation, in this embodiment, a telephone platform is used to process the incoming call of the target user regarding the target case report, and the target user is transferred to the corresponding AI customer service; the relevant call voice is obtained through the call process between the AI ​​customer service and the target user regarding the target case; wherein, the specific implementation process is roughly as follows:

[0064] 1. Phone platform answering and routing

[0065] When the target user dials the insurance company's reporting phone number, the telephone platform automatically answers the call first;

[0066] Using an interactive voice response (IVR) system to guide target users through simple case reporting options, such as "Press 1 to file a vehicle insurance case";

[0067] Based on the target user's selection, the IVR system routes the call to the corresponding AI customer service queue to wait for reception;

[0068] Among them, the IVR system is a telephone automatic service technology that interacts with users through pre-recorded voice prompts or text-to-speech conversion technology, allowing users to select services, obtain information or perform tasks through telephone buttons or voice commands, thereby realizing automated customer service and information processing, improving response efficiency and reducing dependence on manual customer service;

[0069] 2. AI customer service intelligent reception

[0070] When the AI ​​customer service receives a call from the IVR system, it automatically answers the call and establishes a call connection with the target user;

[0071] AI customer service uses the preset automatic speech recognition (ASR) technology to understand the target user's reporting intention and uses natural language processing (NLP) technology to interact with the user by voice;

[0072] 3. Acquisition of call audio

[0073] During the conversation between the AI ​​customer service and the target user regarding the target case, the conversation content is recorded in real time to obtain the conversation audio;

[0074] Through the above process, the target user's call audio regarding the target case can be obtained.

[0075] Furthermore, in the case fraud identification method, the step S200 of converting the call speech into a corresponding target text includes:

[0076] Through NLP speech conversion technology, the call speech is converted into the target text corresponding to the target case.

[0077] In the specific implementation, in this embodiment, the call voice is converted into the target text corresponding to the target case through NLP voice conversion technology; wherein the specific implementation process is roughly as follows:

[0078] Natural language processing (NLP) speech conversion technology is an artificial intelligence application that converts human speech into text information. It captures and processes speech signals through automatic speech recognition (ASR), combined with natural language understanding and generation technology in NLP, to achieve accurate conversion of speech to text, so that machines can understand, respond to and learn human language. It is widely used in many fields such as intelligent assistants, automatic translation, voice command control, etc.

[0079] 1. Call Voice Collection

[0080] When the AI ​​customer service agent is talking to the target user, the conversation content will be automatically recorded to obtain the target case’s conversation voice; the conversation voice data will be securely stored for subsequent processing;

[0081] 2. Speech to Text (ASR)

[0082] Use automatic speech recognition technology to convert the audio files of recorded call speech into preliminary text; the ASR engine can handle different accents and dialects to improve the accuracy of the conversion;

[0083] The conversion process may include preprocessing steps such as noise reduction, speech segmentation, and feature extraction to improve the recognition rate;

[0084] 3. Text optimization and post-processing

[0085] Post-process the preliminary text converted by ASR, including grammar correction, punctuation addition, homophone correction, etc., to improve the readability and accuracy of the text;

[0086] Use NLP technology to understand semantics and extract key information, such as case type, user needs, accident description, etc.

[0087] 4. Extraction and structuring of key information

[0088] Through entity recognition and relationship extraction technology in NLP, key information such as time, place, people, events, etc. are extracted from the processed preliminary text and structured into a format that can be stored in the database;

[0089] 5. Case text creation and integration

[0090] Integrate the extracted key information into a complete case text (i.e., target text) to form a case report or record;

[0091] The generated case text is stored in the case management system together with the relevant call recordings and original audio files for further claims processing and case analysis;

[0092] Through the above process, the content of the user's report call can be efficiently and accurately converted into text form, which is convenient for storage, query and further case processing; at the same time, this also provides technical support for improving customer service quality and claims decision-making efficiency.

[0093] Furthermore, in the case fraud identification method, the step S300, based on the target text, determines whether the target case has a fraud risk by using a preset fraud identification model, comprises:

[0094] Pre-built fraud detection models;

[0095] Inputting the target text into the fraud identification model to generate a fraud identification result of the target case;

[0096] Based on the fraud identification result, determine whether there is a fraud risk in the target case.

[0097] Furthermore, in the case fraud identification method, the pre-constructed fraud identification model includes:

[0098] Collect some user case information;

[0099] After data cleaning and feature selection of the user case information, a corresponding target data set is generated;

[0100] After training and verifying the preset machine learning model using the target data set, a corresponding fraud identification model is generated.

[0101] Furthermore, the case fraud identification method, wherein judging whether the target case has a fraud risk based on the fraud identification result, further includes:

[0102] If there is a risk of fraud in the target case, the target case will be sent to the corresponding claims seat for processing as an abnormal case.

[0103] In the specific implementation, in this embodiment, several user case information is collected; after data cleaning and feature selection of the several user case information, a corresponding target data set is generated; after training and verifying the preset machine learning model using the target data set, a corresponding fraud identification model is generated; the target text is input into the fraud identification model to generate a fraud identification result of the target case; according to the fraud identification result, it is determined whether the target case has a fraud risk; if the target case has a fraud risk, the target case is sent as an abnormal case to the corresponding claims seat for processing; wherein, the specific implementation process is roughly as follows:

[0104] 1. User case information collection

[0105] Collect certain user case information through channels such as phone calls, online forms, and mobile applications, including but not limited to vehicle information, policy details, accident descriptions, loss estimates, etc.;

[0106] 2. Data cleaning and feature selection

[0107] Perform data cleaning on the collected user case information, including processing missing values, outliers, duplicate records, and removing irrelevant features;

[0108] Perform feature selection, using statistical tests, feature importance assessment and other methods to screen out the features that have the greatest impact on fraud identification and build the target data set;

[0109] 3. Machine learning model training and verification

[0110] Use the target dataset and select appropriate machine learning algorithms (such as XGBoost, Catboost, etc.) to build a fraud recognition model;

[0111] The target data set is divided into a training set and a test set. The model is trained using the training set, and the model performance is verified using the test set. The model parameters are adjusted to optimize the recognition accuracy.

[0112] Among them, XGBoost and Catboost are both advanced machine learning algorithms based on the Gradient Boosting framework, which are specially designed to improve the performance of predictive models, especially when dealing with large-scale data sets;

[0113] 4. Application of fraud identification model

[0114] Input the target text into the trained fraud detection model;

[0115] The fraud detection model analyzes the text content of the target text and, combined with the patterns learned during training, generates fraud detection results for the target case, which usually include fraud probability or classification labels;

[0116] 5. Risk assessment and case handling

[0117] Based on the fraud identification results output by the fraud identification model, determine whether the target case has fraud risk through the set threshold;

[0118] If the target case is judged to have fraud risk, it will be marked as an abnormal case and automatically sent to the corresponding claims seat for further investigation and processing by professionals;

[0119] Through the above process, fraud risks can be efficiently identified and managed while ensuring that the claims experience of normal users is not affected, significantly improving the automation level and overall efficiency of the claims process.

[0120] Furthermore, the case fraud identification method, wherein if the target case has a fraud risk, after sending the target case as an abnormal case to a corresponding claims seat for processing, further comprises:

[0121] In response to the broadcast instruction of the call voice by the claims agent, the claims agent manually reviews the target case;

[0122] When the review result shows that the target case is indeed an abnormal case, an investigation instruction is sent to the corresponding investigator, so that the investigator conducts an on-site investigation of the target case.

[0123] In specific implementation, in this embodiment, in response to the broadcast instruction of the call voice by the claims agent, the claims agent manually reviews the target case; when the review result shows that the target case is indeed an abnormal case, an investigation instruction is sent to the corresponding investigator, so that the investigator conducts an on-site investigation of the target case; wherein, the specific implementation process is roughly as follows:

[0124] 1. Broadcast of call voice

[0125] After the claims agent receives the pushed abnormal case (target case) processing request, the claims agent broadcasts the call voice corresponding to the abnormal case to collect necessary information and conduct preliminary evaluation of the case;

[0126] 2. Manual review by claims agents

[0127] The claims agent listens to the call recording and conducts manual review based on the case information; this step may include verification of case details, understanding of the user's situation, and preliminary risk assessment;

[0128] 3. Record and evaluate the review results

[0129] After the review is completed, the claims agent will record the review results and evaluate them; if the case is found to be abnormal or inconsistent with the normal situation, it will be determined as an abnormal case;

[0130] 4. Confirmation of abnormal cases and sending of investigation instructions

[0131] When the review results confirm that the target case is an abnormal case, the process of sending the investigation instruction is automatically triggered; the investigation instruction is sent to the corresponding investigator, including case details, investigation requirements and time limit, etc.;

[0132] 5. On-site investigation and execution

[0133] After receiving the inspection order, the inspector will go to the scene of the target case for on-site inspection according to the information in the order; the inspector will collect on-site evidence, take photos, record accident details, and communicate with the parties involved;

[0134] 6. Feedback on survey results

[0135] After the on-site investigation is completed, the investigator will report the investigation results and the collected evidence; this information will be used for further case analysis and claims decision-making;

[0136] 7. Claims decision and user notification

[0137] Based on the results of the on-site investigation, the claims agent makes a claims decision and notifies the target user of the final result. If further investigation or additional materials are required, they will also be notified to the target user at this time.

[0138] Through the above process, insurance companies can ensure timely and accurate manual review of abnormal cases, while improving the efficiency and quality of the claims process.

[0139] It can be seen from the above method embodiments that the case fraud identification method provided by the present invention at least includes: obtaining the call voice of the target user about the target case; converting the call voice into the corresponding target text; based on the target text, judging whether the target case has a fraud risk through a preset fraud identification model; etc. In this way, the method of the present invention can automatically identify whether the user case has a fraud risk, improve the claims efficiency, and improve the user experience.

[0140] Specifically, the case fraud identification method provided by the embodiment of the present invention is that when a user reports a case, the telephone platform is responsible for real-time transfer, and then the user's voice is converted into text in real time through NLP, and then the case is identified for fraud, abnormal cases that may involve fraud are identified, and finally the case is transferred to the claims seat for processing; the method of the present invention identifies whether there is a risk of fraud in the user's case and pre-processes the possible fraud risk, thereby effectively improving the case processing efficiency of the claims seat.

[0141] It should be understood that, although the present application provides method operation steps as described in the embodiments or flowcharts, more or less operation steps may be included based on conventional or non-creative labor, and these operation steps are not necessarily performed in sequence according to the embodiment or flowchart. The order of steps listed in the embodiment or flowchart is only one way of executing the order of many steps, and does not represent the only execution order. It should be noted that there is not necessarily a certain order between the above steps. A person of ordinary skill in the art can understand from the description of the embodiment of the present invention that in different embodiments, the above steps may have different execution orders, that is, they may be executed in parallel, or they may be executed in exchange, etc. Moreover, at least a part of the steps in the embodiment or flowchart may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but may be executed at different times, and the execution order of these sub-steps or stages is not necessarily performed in sequence, but may be executed in turn, alternately or synchronously with other steps or at least a part of the sub-steps or stages of other steps.

[0142] Based on the above method embodiment, please refer to Figure 2 Another embodiment of the present invention further provides a case fraud identification system, wherein the system comprises:

[0143] An acquisition module 11 is used to acquire the target user's call voice regarding the target case;

[0144] A conversion module 12, used for converting the call speech into a corresponding target text;

[0145] The judgment module 13 is used to judge whether there is a fraud risk in the target case based on the target text through a preset fraud recognition model.

[0146] Furthermore, in the case fraud identification system, the step of obtaining the target user's call voice regarding the target case includes:

[0147] Processing incoming calls from target users regarding reporting target cases through a telephone platform, and transferring the target users to corresponding AI customer service;

[0148] Relevant call audio is obtained through the call process between the AI ​​customer service and the target user regarding the target case.

[0149] In the specific implementation, in this embodiment, a telephone platform is used to process the incoming call of the target user regarding the target case report, and the target user is transferred to the corresponding AI customer service; the relevant call voice is obtained through the call process between the AI ​​customer service and the target user regarding the target case; wherein, the specific implementation process is roughly as follows:

[0150] 1. Phone platform answering and routing

[0151] When the target user dials the insurance company's reporting phone number, the telephone platform automatically answers the call first;

[0152] Using an interactive voice response (IVR) system to guide target users through simple case reporting options, such as "Press 1 to file a vehicle insurance case";

[0153] Based on the target user's selection, the IVR system routes the call to the corresponding AI customer service queue to wait for reception;

[0154] Among them, the IVR system is a telephone automatic service technology that interacts with users through pre-recorded voice prompts or text-to-speech conversion technology, allowing users to select services, obtain information or perform tasks through telephone buttons or voice commands, thereby realizing automated customer service and information processing, improving response efficiency and reducing dependence on manual customer service;

[0155] 2. AI customer service intelligent reception

[0156] When the AI ​​customer service receives a call from the IVR system, it automatically answers the call and establishes a call connection with the target user;

[0157] AI customer service uses the preset automatic speech recognition (ASR) technology to understand the target user's reporting intention and uses natural language processing (NLP) technology to interact with the user by voice;

[0158] 3. Acquisition of call audio

[0159] During the conversation between the AI ​​customer service and the target user regarding the target case, the conversation content is recorded in real time to obtain the conversation audio;

[0160] Through the above process, the target user's call audio regarding the target case can be obtained.

[0161] Furthermore, in the case fraud identification system, the step of converting the call speech into a corresponding target text includes:

[0162] Through NLP speech conversion technology, the call speech is converted into the target text corresponding to the target case.

[0163] In the specific implementation, in this embodiment, the call voice is converted into the target text corresponding to the target case through NLP voice conversion technology; wherein the specific implementation process is roughly as follows:

[0164] Natural language processing (NLP) speech conversion technology is an artificial intelligence application that converts human speech into text information. It captures and processes speech signals through automatic speech recognition (ASR), combined with natural language understanding and generation technology in NLP, to achieve accurate conversion of speech to text, so that machines can understand, respond to and learn human language. It is widely used in many fields such as intelligent assistants, automatic translation, voice command control, etc.

[0165] 1. Call Voice Collection

[0166] When the AI ​​customer service agent is talking to the target user, the conversation content will be automatically recorded to obtain the target case’s conversation voice; the conversation voice data will be securely stored for subsequent processing;

[0167] 2. Speech to Text (ASR)

[0168] Use automatic speech recognition technology to convert the audio files of recorded call speech into preliminary text; the ASR engine can handle different accents and dialects to improve the accuracy of the conversion;

[0169] The conversion process may include preprocessing steps such as noise reduction, speech segmentation, and feature extraction to improve the recognition rate;

[0170] 3. Text optimization and post-processing

[0171] Post-process the preliminary text converted by ASR, including grammar correction, punctuation addition, homophone correction, etc., to improve the readability and accuracy of the text;

[0172] Use NLP technology to understand semantics and extract key information, such as case type, user needs, accident description, etc.

[0173] 4. Extraction and structuring of key information

[0174] Through entity recognition and relationship extraction technology in NLP, key information such as time, place, people, events, etc. are extracted from the processed preliminary text and structured into a format that can be stored in the database;

[0175] 5. Case text creation and integration

[0176] Integrate the extracted key information into a complete case text (i.e., target text) to form a case report or record;

[0177] The generated case text is stored in the case management system together with the relevant call recordings and original audio files for further claims processing and case analysis;

[0178] Through the above process, the content of the user's report call can be efficiently and accurately converted into text form, which is convenient for storage, query and further case processing; at the same time, this also provides technical support for improving customer service quality and claims decision-making efficiency.

[0179] Furthermore, the case fraud identification system, wherein the step of judging whether the target case has a fraud risk based on the target text by using a preset fraud identification model, includes:

[0180] Pre-built fraud detection models;

[0181] Inputting the target text into the fraud identification model to generate a fraud identification result of the target case;

[0182] Based on the fraud identification result, determine whether there is a fraud risk in the target case.

[0183] Furthermore, in the case fraud identification system, the pre-built fraud identification model includes:

[0184] Collect some user case information;

[0185] After data cleaning and feature selection of the user case information, a corresponding target data set is generated;

[0186] After training and verifying the preset machine learning model using the target data set, a corresponding fraud identification model is generated.

[0187] Furthermore, the case fraud identification system, wherein judging whether the target case has a fraud risk based on the fraud identification result, further includes:

[0188] If there is a risk of fraud in the target case, the target case will be sent to the corresponding claims seat for processing as an abnormal case.

[0189] In the specific implementation, in this embodiment, several user case information is collected; after data cleaning and feature selection of the several user case information, a corresponding target data set is generated; after training and verifying the preset machine learning model using the target data set, a corresponding fraud identification model is generated; the target text is input into the fraud identification model to generate a fraud identification result of the target case; according to the fraud identification result, it is determined whether the target case has a fraud risk; if the target case has a fraud risk, the target case is sent to the corresponding claims seat as an abnormal case for processing; wherein, the specific implementation process is roughly as follows:

[0190] 1. User case information collection

[0191] Collect certain user case information through channels such as phone calls, online forms, and mobile applications, including but not limited to vehicle information, policy details, accident descriptions, loss estimates, etc.;

[0192] 2. Data cleaning and feature selection

[0193] Perform data cleaning on the collected user case information, including processing missing values, outliers, duplicate records, and removing irrelevant features;

[0194] Perform feature selection, using statistical tests, feature importance assessment and other methods to screen out the features that have the greatest impact on fraud identification and build the target data set;

[0195] 3. Machine learning model training and verification

[0196] Use the target dataset and select appropriate machine learning algorithms (such as XGBoost, Catboost, etc.) to build a fraud recognition model;

[0197] The target data set is divided into a training set and a test set. The model is trained using the training set, and the model performance is verified using the test set. The model parameters are adjusted to optimize the recognition accuracy.

[0198] Among them, XGBoost and Catboost are both advanced machine learning algorithms based on the Gradient Boosting framework, which are specially designed to improve the performance of predictive models, especially when dealing with large-scale data sets;

[0199] 4. Application of fraud identification model

[0200] Input the target text into the trained fraud detection model;

[0201] The fraud detection model analyzes the text content of the target text and, combined with the patterns learned during training, generates fraud detection results for the target case, which usually include fraud probability or classification labels;

[0202] 5. Risk assessment and case handling

[0203] Based on the fraud identification results output by the fraud identification model, determine whether the target case has fraud risk through the set threshold;

[0204] If the target case is judged to have fraud risk, it will be marked as an abnormal case and automatically sent to the corresponding claims seat for further investigation and processing by professionals;

[0205] Through the above process, fraud risks can be efficiently identified and managed while ensuring that the claims experience of normal users is not affected, significantly improving the automation level and overall efficiency of the claims process.

[0206] Furthermore, the case fraud identification system, wherein if the target case has a fraud risk, after sending the target case as an abnormal case to a corresponding claims seat for processing, further includes:

[0207] In response to the broadcast instruction of the call voice by the claims agent, the claims agent manually reviews the target case;

[0208] When the review result shows that the target case is indeed an abnormal case, an investigation instruction is sent to the corresponding investigator, so that the investigator conducts an on-site investigation of the target case.

[0209] In specific implementation, in this embodiment, in response to the broadcast instruction of the call voice by the claims agent, the claims agent manually reviews the target case; when the review result shows that the target case is indeed an abnormal case, an investigation instruction is sent to the corresponding investigator, so that the investigator conducts an on-site investigation of the target case; wherein, the specific implementation process is roughly as follows:

[0210] 1. Broadcast of call voice

[0211] After the claims agent receives the pushed abnormal case (target case) processing request, the claims agent broadcasts the call voice corresponding to the abnormal case to collect necessary information and conduct preliminary evaluation of the case;

[0212] 2. Manual review by claims agents

[0213] The claims agent listens to the call recording and conducts manual review based on the case information; this step may include verification of case details, understanding of the user's situation, and preliminary risk assessment;

[0214] 3. Record and evaluate the review results

[0215] After the review is completed, the claims agent will record the review results and evaluate them; if the case is found to be abnormal or inconsistent with the normal situation, it will be determined as an abnormal case;

[0216] 4. Confirmation of abnormal cases and sending of investigation instructions

[0217] When the review results confirm that the target case is an abnormal case, the process of sending the investigation instruction is automatically triggered; the investigation instruction is sent to the corresponding investigator, including case details, investigation requirements and time limit, etc.;

[0218] 5. On-site investigation and execution

[0219] After receiving the inspection order, the inspector will go to the scene of the target case for on-site inspection according to the information in the order; the inspector will collect on-site evidence, take photos, record accident details, and communicate with the parties involved;

[0220] 6. Feedback on survey results

[0221] After the on-site investigation is completed, the investigator will report the investigation results and the collected evidence; this information will be used for further case analysis and claims decision-making;

[0222] 7. Claims decision and user notification

[0223] Based on the results of the on-site investigation, the claims agent makes a claims decision and notifies the target user of the final result. If further investigation or additional materials are required, they will also be notified to the target user at this time.

[0224] Through the above process, insurance companies can ensure timely and accurate manual review of abnormal cases, while improving the efficiency and quality of the claims process.

[0225] It can be seen from the above system embodiments that the case fraud identification system provided by the present invention at least includes: obtaining the call voice of the target user about the target case; converting the call voice into the corresponding target text; based on the target text, judging whether the target case has a fraud risk through a preset fraud identification model; etc. In this way, the system of the present invention can automatically identify whether the user's case has a fraud risk, improve the claims efficiency, and improve the user experience.

[0226] Specifically, in the case fraud identification system provided by the embodiment of the present invention, when a user calls in to report a case, the telephone platform is responsible for real-time transfer, and then the user's voice is converted into text in real time through NLP, and then the case is identified for fraud, abnormal cases that may involve fraud are identified, and finally the case is transferred to the claims seat for processing; the system of the present invention identifies whether there is a risk of fraud in the user's case and pre-processes the possible fraud risk, thereby effectively improving the case processing efficiency of the claims seat.

[0227] Based on the above method embodiment, please refer to Figure 3 Another embodiment of the present invention further provides a computer device, wherein the computer device 10 comprises:

[0228] memory 120 and one or more processors 110, Figure 3 A processor 110 is used as an example for description. The processor 110 and the memory 120 may be connected via a communication bus or other means. Figure 3 The example of connecting via a communication bus is taken as an example.

[0229] The processor 110 is used to complete various control logics of the computer device 10. It can be a general-purpose processor, a digital signal processor (DSP), an application-specific integrated circuit (ASIC), a field programmable gate array (FPGA), a single-chip microcomputer, an ARM (Acorn RISC Machine) or other programmable logic device, a discrete gate or transistor logic, a discrete hardware component, or any combination of these components. In addition, the processor 110 can also be any traditional processor, microprocessor or state machine. The processor 110 can also be implemented as a combination of computing devices, for example, a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors combined with a DSP, or any other such configuration.

[0230] The memory 120 is a non-volatile computer-readable storage medium that can be used to store non-volatile software programs, non-volatile computer executable programs and modules, such as the computer program corresponding to the case fraud identification method in the embodiment of the present invention. The processor 110 executes various functional applications and data processing of the computer device 10 by running the non-volatile software programs, instructions and units stored in the memory 120, that is, implements the case fraud identification method in the above method embodiment.

[0231] The memory 120 may include a program storage area and a data storage area, wherein the program storage area may store an application required for operating the device or at least one function; the data storage area may store data created according to the use of the computer device 10, etc. In addition, the memory 120 may include a high-speed random access memory, and may also include a non-volatile memory, such as at least one disk storage device, a flash memory device, or other non-volatile solid-state storage device. In some embodiments, the memory 120 may optionally include a memory remotely arranged relative to the processor 110, and these remote memories may be connected to the computer device 10 via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0232] One or more units are stored in the memory 120, and when executed by one or more processors 110, the case fraud identification method in any of the above method embodiments can be implemented, for example, the above described method can be implemented. Figure 1 The method comprises steps S100 to S300.

[0233] Those skilled in the art will understand that Figure 3 The hardware structure diagram shown in the figure is only a schematic diagram of a partial structure related to the solution of the present invention, and does not constitute a limitation on the computer device to which the solution of the present invention is applied. The specific computer device may include more components than those shown in the figure, or combine certain components, or have a different arrangement of components.

[0234] Based on the above method embodiments, another embodiment of the present invention further provides a non-volatile computer-readable storage medium, wherein the non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by at least one processor, the case fraud identification method in any of the above method embodiments can be implemented, for example, the above described method can be implemented. Figure 1 The method comprises steps S100 to S300.

[0235] As an example, the non-volatile storage medium can include a read-only memory (ROM), a programmable ROM (PROM), an electrically programmable ROM (EPROM), an electrically erasable ROM (EEPROM), or a flash memory. Volatile memory can include a random access memory (RAM) as an external cache memory. By way of illustration and not limitation, RAM can be obtained in many forms such as synchronous RAM (SRAM), dynamic RAM, (DRAM), synchronous DRAM (SDRAM), double data rate SDRAM (DDR SDRAM), enhanced SDRAM (ESDRAM), Synch link DRAM (SLDRAM), and direct Rambus RAM (DRRAM). The disclosed memory components or memories of the operating environment described herein are intended to include one or more of these and / or any other suitable types of memory.

[0236] Based on the above method embodiments, another embodiment of the present invention further provides a computer program product, the computer program product includes a computer program stored on a non-volatile computer-readable storage medium, the computer program includes program instructions, when the program instructions are executed by a processor, the case fraud identification method in any of the above method embodiments can be implemented, for example, the above described method can be implemented Figure 1 The method comprises steps S100 to S300.

[0237] The embodiments described above are merely illustrative, in which the units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0238] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a general hardware platform, and of course can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the relevant technology can be embodied in the form of a software product, and the computer software product can exist in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods of each embodiment or some parts of the embodiment.

[0239] Conditional language such as "can," "can," "might," or "may," unless specifically stated otherwise or otherwise understood within the context as used, is generally intended to convey, among other things, that a particular embodiment can include (while other embodiments do not) a particular feature, element, and / or operation. Thus, such conditional language is also generally intended to imply that a feature, element, and / or operation is required for one or more embodiments anyway or that one or more embodiments must include logic for determining, with or without input or prompting, whether such features, elements, and / or operations are included or will be performed in any particular embodiment.

[0240] What has been described herein, in this specification and in the drawings includes examples that can provide a case fraud identification method, system, computer device, and non-volatile computer-readable storage medium. Of course, it is not possible to describe every conceivable combination of elements and / or methods for the purpose of describing the various features of the present disclosure, but it can be recognized that many other combinations and permutations of the disclosed features are possible. Therefore, it is obvious that various modifications can be made to the present disclosure without departing from the scope or spirit of the present disclosure, but all these various modifications should belong to the scope of protection of the claims attached to the present invention. In addition, or in an alternative, other embodiments of the present disclosure may be obvious from consideration of the present specification and the drawings and from the practice of the present disclosure as presented herein. It is intended that the examples proposed in the present specification and the drawings are considered to be illustrative and not restrictive in all respects. Although specific terms are used herein, they are used in a general and descriptive sense and are not used for limiting purposes.

Claims

1. A case fraud identification method, characterized in that: include: Obtain the target user's call audio regarding the target case; Converting the call speech into corresponding target text; Based on the target text, a preset fraud identification model is used to determine whether the target case has a fraud risk.

2. The case fraud identification method according to claim 1, characterized in that: The obtaining of the target user's call voice regarding the target case includes: Processing incoming calls from target users regarding reporting target cases through a telephone platform, and transferring the target users to corresponding AI customer service; Relevant call audio is obtained through the call process between the AI ​​customer service and the target user regarding the target case.

3. The case fraud identification method according to claim 2, characterized in that: The converting the call speech into a corresponding target text includes: Through NLP speech conversion technology, the call speech is converted into the target text corresponding to the target case.

4. The case fraud identification method according to claim 3 is characterized in that: The step of judging whether the target case has a fraud risk based on the target text by using a preset fraud identification model includes: Pre-built fraud detection models; Inputting the target text into the fraud identification model to generate a fraud identification result of the target case; Based on the fraud identification result, determine whether there is a fraud risk in the target case.

5. The case fraud identification method according to claim 4 is characterized in that: The pre-built fraud identification model includes: Collect some user case information; After data cleaning and feature selection of the user case information, a corresponding target data set is generated; After training and verifying the preset machine learning model using the target data set, a corresponding fraud identification model is generated.

6. The case fraud identification method according to claim 5, characterized in that: The determining, based on the fraud identification result, whether the target case has a fraud risk further includes: If there is a risk of fraud in the target case, the target case will be sent to the corresponding claims seat for processing as an abnormal case.

7. The case fraud identification method according to claim 6, characterized in that: If the target case has a fraud risk, the target case is sent as an abnormal case to a corresponding claims seat for processing, and further includes: In response to the broadcast instruction of the call voice by the claims agent, the claims agent manually reviews the target case; When the review result shows that the target case is indeed an abnormal case, an investigation instruction is sent to the corresponding investigator, so that the investigator conducts an on-site investigation of the target case.

8. A case fraud identification system, characterized in that: include: An acquisition module is used to acquire the target user's call voice regarding the target case; A conversion module, used for converting the call speech into a corresponding target text; The judgment module is used to judge whether there is a fraud risk in the target case based on the target text through a preset fraud recognition model.

9. A computer device, characterized in that: The computer device comprises at least one processor; and, a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program that can be executed by the at least one processor. When the computer program is executed by the at least one processor, the case fraud identification method as described in any one of claims 1-7 can be implemented.

10. A non-volatile computer-readable storage medium, characterized in that: The non-volatile computer-readable storage medium stores a computer program, and when the computer program is executed by at least one processor, it can implement the case fraud identification method as described in any one of claims 1-7.

Citation Information

Patent Citations

  • Risk identification method and device for vehicle insurance claim settlement case, equipment and storage medium

    CN109919783A

  • Method and device for identifying insurance claim fraudulent behavior

    CN112800272A