Disclosure of Invention
The invention mainly aims to provide an intelligent auditing method, device, equipment and storage medium for insurance data, and aims to solve the technical problems that in the existing insurance business, the time consumption of manual input and auditing is long, and the auditing is easily affected by human errors and negligence, so that the auditing speed of an insurance business bill is low and the data accuracy is insufficient.
In order to achieve the above object, the present invention provides an intelligent auditing method for insurance data, comprising:
acquiring customer verification data, wherein the customer verification data comprises an identity card picture of a customer and a target insurance certificate picture, and is used for auditing insurance handling data;
identifying the customer verification data based on a pre-trained identification model, and generating a verification data identification result, wherein the verification data identification result comprises target insurance certificate information;
extracting audit point information from the verification data identification result according to a predefined audit standard;
Comparing and analyzing the audit point information with corresponding information in the insurance transaction information, and judging whether the information in the insurance transaction information is consistent with the information in the customer verification information;
and if the comparison and analysis result is that the information is consistent, confirming that the insurance transaction data passes the audit.
In one embodiment, after determining whether the information in the insurance transaction material is consistent with the information in the customer verification material, the method further comprises:
if the comparison and analysis result is that the information is inconsistent, confirming that the insurance transaction data does not pass the audit;
identifying and determining a field and a corresponding position which are inconsistent with the audit point information in the insurance transaction information;
Generating audit prompt information based on the inconsistent fields and the corresponding positions, wherein the audit prompt information comprises data items needing to be re-submitted or to be submitted in a supplementing mode;
And sending the auditing prompt information to a salesman end to prompt resubmit or supplement and submit auditing point information which does not pass auditing in the insurance transaction information.
In one embodiment, obtaining customer verification material includes:
Generating an H5 data uploading short link, wherein the uploading short link comprises a URL of an uploading page and parameters for identifying client identity information and data types;
The uploading short link is sent to the mobile terminal of the client through a short message;
When the uploading short link is triggered, providing an uploading page, and displaying corresponding data uploading options according to the data type;
Receiving the identity card picture and the target insurance certificate picture uploaded by the data uploading option as the customer verification data;
checking whether the client verification data meets the format requirement;
And storing the client verification data meeting the format requirements into a database, and associating the client verification data with the client identity information.
In one embodiment, before identifying the customer verification material based on the pre-trained identification model, further comprising:
acquiring sample data, and dividing the sample data into a training set and a testing set;
Inputting the training set into a deep learning model for training, and adjusting the recognition parameters of the deep learning model root by learning the data characteristics in the training set to generate a preliminary recognition model;
inputting the test set into the preliminary identification model to obtain a test result;
analyzing the recognition accuracy of the preliminary recognition model based on the test result, and adjusting the recognition parameters of the preliminary recognition model according to the recognition accuracy;
And training the preliminary recognition model for multiple times by using the test set until the recognition accuracy reaches an accuracy threshold, stopping training, and generating a final recognition model.
In one embodiment, extracting audit point information from the verification material identification results according to predefined audit criteria includes:
Defining audit criteria for the insurance transaction data based on target insurance business requirements;
Corresponding analysis is carried out according to the fields in the auditing standard and the fields in the identification result of the verification data, and auditing point information needing to be extracted is determined;
and extracting the audit point information, and formatting the audit point information based on the format standard of the insurance transaction information.
In one embodiment, comparing the audit point information with corresponding information in the insurance transaction information to determine whether the information in the insurance transaction information is consistent with the information in the customer verification information includes:
Extracting a data field matched with the type of the audit point information from the insurance transaction information as information to be audited;
comparing each data field of the information to be checked with the corresponding field in the information of the checking point one by one;
If each field in the information to be checked is completely consistent with the corresponding field in the information of the checking point, confirming that the comparison and analysis result is consistent with the information; if the data field inconsistent with the audit point information exists in the information to be audited, confirming that the comparison and analysis result is inconsistent with the information;
When the comparison analysis is completed, a comparison analysis result report is generated, and the comparison result of each field is recorded.
In one embodiment, after confirming that the insurance transaction material passes the audit, further comprising:
Determining the service type of the insurance transaction data, and determining a service bill template based on the service type;
Filling the information of the insurance transaction information into the service list template to generate an insurance service list;
checking whether the information in the insurance service ticket is consistent with the corresponding audit point information;
And storing the verified and consistent insurance business list into an insurance archive.
Further, in order to achieve the above object, the present invention also provides an intelligent audit device for insurance data, where the intelligent audit device for insurance data includes a memory, a processor, and an intelligent audit program for insurance data stored in the memory and capable of running on the processor, where the intelligent audit program for insurance data implements the steps of the intelligent audit method for insurance data described above when executed by the processor.
Further, in order to achieve the above object, the present invention also provides a computer storage medium, on which an insurance data intelligent auditing program is stored, which when executed by a processor, implements the steps of the insurance data intelligent auditing method as described above.
The invention relates to an intelligent auditing method for insurance data, which comprises the steps of acquiring customer verification data, wherein the data comprises an identity card picture and a target insurance certificate picture and is used for auditing insurance transaction data. And identifying the customer verification data by using the pre-trained identification model to generate a verification data identification result containing the target insurance certificate information. And extracting audit point information from the verification data identification result according to a predefined audit standard. And comparing and analyzing the audit point information with the corresponding information in the insurance transaction information to judge whether the information in the insurance transaction information is consistent with the information in the customer verification information. If the comparison and analysis result shows that the information is consistent, confirming that the insurance transaction data passes the audit. Through automatic processing and intelligent recognition, the invention improves auditing efficiency, reduces manual intervention and error rate, thereby accelerating the processing speed of insurance business orders and improving customer experience and business processing accuracy.
Detailed Description
It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
In the conventional car insurance business application process, a customer needs to provide a series of related information to complete an insurance application. Such information includes, but is not limited to, vehicle information, driver license information, and personal identification. The customer typically needs to submit this information to the insurance company's business personnel, who will enter the data provided by the customer into the order system to generate the insurance policy.
The business personnel manually fills out the insurance business form according to the data provided by the clients. This process includes entering various customer information and insurance details into the service ticket. Errors may occur in the manual filling process, such as information entry omission, data input errors or inconsistent formats.
Insurance companies typically audit completed insurance policies by underwriters. The underwriter will check whether the information in the service ticket is complete and accurate. Manual auditing typically includes checking customer information and insurance clauses item by item to ensure accuracy and compliance of the business form.
The manual auditing process is time consuming because the underwriter needs to check each data point of the insurance policy item by item. This piece-by-piece auditing approach results in a slower overall auditing process. Because the time of manual auditing is longer, the generation and auditing of the insurance business ticket are also affected, thereby prolonging the waiting time of the client.
The slow auditing process may result in excessive waiting time of the client after applying insurance, which affects the client experience. The long waiting time and uncertain auditing process of the customer may lead to an unsatisfactory user experience when applying for car insurance.
In modern fast paced society, customers desire to be able to complete insurance applications and auditing processes quickly, and existing manual auditing approaches fail to meet this need.
Referring to fig. 1, fig. 1 is a flowchart of an embodiment of an intelligent audit method for insurance data according to the present invention. It should be noted that although a logical order is depicted in the flowchart, in some cases the steps depicted or described may be performed in a different order than presented herein.
As shown in FIG. 1, the intelligent auditing method of insurance data provided by the invention comprises the following steps:
S10, acquiring customer verification data, wherein the customer verification data comprises an identity card picture of a customer and a target insurance certificate picture, and is used for auditing insurance transaction data;
in this embodiment, verification data for auditing is obtained from a customer, wherein the verification data includes an identification card picture of the customer and a target insurance certificate picture. These materials are used to confirm the authenticity and integrity of the information provided by the customer.
The target insurance certificate is a verification document directly related to the insurance target and is used for confirming the validity and detailed information of the insurance target. Specifically, it includes formal proof files for use in insurance business, which record specific information of insurance contract and serve as important credentials in application, claim or management process. The main function of the target insurance certificate is to provide the legal and effective proof of insurance mark and ensure the authenticity and integrity of insurance business. For example, the target insurance certificate may be a driver license, an official document for proving personal driving qualification. The method records the basic information of the driver and the driving permission thereof, and the content comprises a driving license number, a name of a driver's license, a birth date, an address, a certification authority, a driving category and the like. Or a document for confirming ownership of the vehicle, which may be a certificate of registration of the vehicle owner. It is typically issued by a vehicle registration authority, the content including owner name, vehicle registration number, vehicle make and model, vehicle Identification Number (VIN), registration date, etc. And the file or card is used for proving the coverage of personal health insurance, and the content comprises an insurance policy number, a cardholder name, an insurance company name, an insurance coverage, an expiration date and the like.
When the customer applies for insurance service, the verification data is uploaded through the online platform or mobile application of the insurance company. These platforms should support a variety of picture formats, such as JPEG, PNG. The system provides a user-friendly upload interface including clear indication and upload buttons for clients to upload identification cards and insurance voucher pictures. The system specifies the format (e.g., JPEG or PNG) of the uploaded picture to ensure that the picture can be properly processed and identified. The system sets a file size limit (e.g., no more than 5 MB) to ensure quality of the uploaded pictures and processing efficiency of the system. After receiving the picture uploaded by the client, the system performs preliminary inspection on the picture, including file format, resolution, size and the like, to confirm that the picture meets the system requirements. The pictures meeting the requirements are stored in a database of the system and are associated with other information (such as client ID) of the client, so that the follow-up auditing and processing are convenient. The uploaded verification data can be stored in an encrypted mode, so that the privacy of a client is ensured to be protected, and unauthorized access is prevented. The system meets the requirements of relevant laws and regulations on data protection, and ensures the legitimacy of processing and storing the client data.
In one embodiment, on an online insurance application platform, a customer fills out an application form and uploads a photograph of an identification card and an insurance certificate. The system checks the file format and size, then stores the qualified file in the cloud storage service and associates with the customer's account.
And the verification process of the client is optimized, so that the verification efficiency and accuracy of the insurance data are improved. The system automatically processes and verifies the uploaded identity card and insurance evidence picture, reduces the requirement of manual operation, and reduces the risk caused by human errors.
S20, identifying the customer verification data based on a pre-trained identification model, and generating a verification data identification result, wherein the verification data identification result comprises target insurance certificate information;
In this embodiment, the authentication data uploaded by the client is identified using an already trained identification model. The verification material typically includes an identification card picture and a target insurance certificate picture. The recognition model extracts key information in the image by learning a large amount of data.
Deep learning algorithms (e.g., convolutional Neural Networks (CNNs), transformers, etc.) are used to train to accurately identify and extract information in various verification materials. The model is trained using a training set containing annotation data (including identification cards and insurance evidence pictures under different conditions). The training data includes various samples to enhance the model's ability to adapt to different picture quality and formats. After training is completed, the performance of the model is verified through the test set, and the identification accuracy of the model under different conditions is ensured.
The identity card and the target insurance certificate picture uploaded by the client through the system can be sent into the identification model for processing. Prior to recognition, the system pre-processes the uploaded picture, such as adjusting image resolution, removing noise, and correcting tilt, to ensure that the image quality is appropriate for model recognition.
After the model identifies the uploaded verification data, the generated result contains the target insurance certificate information extracted from the verification data. These recognition results are the basis for subsequent data processing and auditing. Personal information such as name, identification card number, date of birth, etc. is extracted from the identification card picture. For example, key information such as an insurance policy number, an insurance company name, an insurance amount, an insurance validity period, etc. is extracted from the insurance credential picture. And the original data output by the identification model is arranged, so that the information is ensured to be formatted into a standardized data format, and the subsequent data processing and comparison are convenient. And storing the extracted verification data information in a database, and ensuring the integrity and accessibility of the data so as to support the subsequent auditing process.
The client verification data is intelligently identified based on the pre-trained identification model, so that the accuracy and the processing efficiency of information extraction are greatly improved. The automatic identification process reduces the dependence on manual operation, reduces the error rate and accelerates the information processing speed. The method not only improves the overall efficiency of insurance data auditing, but also enhances the accuracy of data processing, optimizes the processing flow of insurance business, and improves the service experience of clients.
S30, according to a predefined auditing standard, auditing point information is extracted from the verification data identification result;
In this embodiment, key audit point information is extracted from the model identification result according to a preset audit standard. The audit point information is a data point for verifying and confirming the accuracy of the insurance business ticket, and comprises key information to be checked in the identity card and the target insurance certificate.
An audit standard is set in the system to define the data fields to be checked. For example, the audit criteria for an identification card may include name, identification card number, and date of birth, while the audit criteria for an insurance voucher may include an insurance number, insurance company name, and insurance amount, among others. And defining a mapping relation between the fields in the auditing standard and the fields in the identification result, and ensuring that each field can be accurately extracted from the identification result. And analyzing and arranging the original data (such as the identity card and the insurance certificate information) obtained from the identification model. The recognition result may include multiple data fields that the system needs to filter and extract according to audit criteria. And extracting key field information meeting auditing standards from the identification result. For example, the name, the number of the identification card, the date of birth, etc. are extracted. The insurance certificate is used for extracting insurance numbers, insurance company names, insurance amounts and the like.
Formatting the extracted audit point information, and arranging the audit point information into a standardized data format so as to facilitate subsequent comparison and verification. And storing the formatted audit point information in a database, and ensuring the integrity and consistency of the data for use in the subsequent audit step.
By extracting the audit point information from the identification result according to the predefined audit standard, the accuracy and consistency of information extraction are improved. The clear auditing standard helps the system to accurately screen out key data from the identification result, and ensures that the data in the insurance business bill meets auditing requirements. Manual intervention and errors are reduced, the data processing flow is optimized, and the auditing speed of the insurance business ticket is accelerated, so that the overall business processing efficiency and the customer experience are improved.
S40, comparing and analyzing the audit point information with corresponding information in the insurance transaction information, and judging whether the information in the insurance transaction information is consistent with the information in the customer verification information;
In this embodiment, the audit point information extracted from the customer verification data is compared with the relevant data in the insurance transaction data to confirm whether the information therebetween is consistent. The comparison and analysis process is a key step in ensuring the accuracy of the insurance policy, and it can verify whether the information provided by the customer meets the requirements in the insurance transaction data. Through systematic comparison analysis, the problem of inconsistent information can be identified and corrected, so that the integrity and reliability of the insurance transaction flow are ensured.
Key information extracted from the customer's verification material (e.g., identification card and insurance certificate picture). Such information typically includes identification numbers, applicant names, insurance policy numbers, and the like. And converting the text data extracted from the pictures uploaded by the clients into a structured data format by using an image recognition model. For example, the name and identification number extracted from the identification card are converted into text fields.
Related data extracted from insurance transaction information (such as insurance application forms, electronic insurance policies, etc.). Such data includes policy number, applicant name, insurance amount, etc. And automatically extracting field data from the insurance handling system and carrying out structural processing. For example, the policy number and the insurance amount are extracted from the electronic policy.
And comparing the extracted audit point information with corresponding fields in the insurance transaction information item by item. Ensuring that each field (e.g., policy number, applicant name) is completely consistent between the two. Character matching techniques, regular expression analysis, or data cleansing algorithms are applied to ensure accurate comparison of information. The problems of format differences, inconsistent character cases, extra spaces, etc. which may exist are handled.
And verifying whether the formats of the data fields are consistent. For example, the policy number should conform to a particular format and the identification card number should be of the correct character length and format. It is checked whether the extracted data is logically consistent. For example, if the applicant's name matches in the identification card and policy, the amount of the policy is in compliance with the policy terms.
The system records the comparison result of each field, including consistent and inconsistent detailed information. For example, the contents, location, and expected and actual values of the inconsistent fields are recorded. A detailed report is generated for each inconsistent field, the reasons and the positions of the inconsistencies are described, and the basis is provided for subsequent processing.
And for the condition that the comparison result shows that the information is inconsistent, marking the auditing state as 'failed'. The system generates notification or prompt information informing the business personnel or auditor of the information items which need to be rechecked or revised.
The comparison of each field is recorded in detail, indicating consistent and inconsistent data. Listing the abnormal conditions found in the comparison process and providing a solution suggestion or a next operation guide. And the data inconsistency problem is further analyzed and processed by the auditing personnel. As part of the system record, it is convenient for future audit and data backtracking.
By comparing and analyzing the information of the auditing points with the corresponding information in the insurance transaction data, the accuracy and efficiency of data auditing are improved. The automatic comparison process of the system reduces manual intervention, reduces the risk of operation errors and accelerates the information processing speed. The method not only improves the accuracy of the insurance transaction data, but also ensures the consistency of the customer verification data and the insurance transaction data, thereby optimizing the auditing flow of insurance business, improving customer experience and improving the efficiency of overall business processing.
And S50, if the comparison and analysis results are consistent, confirming that the insurance transaction data passes the audit.
In this embodiment, the final result is obtained from the comparison and analysis step to determine whether the insurance transaction data passes the audit. The comparison analysis results are the output of a consistency check between the customer verification material and the insurance transaction material. In the alignment analysis step, the system stores the alignment results, including the alignment status (consistent or inconsistent) of each field. And extracting the whole consistency information from the stored comparison result, and preparing for subsequent verification of the auditing state. And assembling the comparison analysis result into a processable format, integrating the comparison states of all the fields, and generating a final auditing decision basis.
And comparing the analysis results to judge the consistency of the analysis results, and confirming whether all the extracted information is completely consistent. This step determines whether to flag the insurance transaction data as passing the audit. The system checks the results of each comparison field to determine if all fields are completely consistent in the security transaction material and the customer verification material. If the comparison results of all the fields are consistent, judging that the information is consistent as a whole, otherwise, recording inconsistent information and marking that the information fails to pass the audit. Predefined rules and algorithms, such as character matching algorithms or data verification techniques, are applied to ensure complete agreement of the information. The comparison results of all the fields are integrated to obtain the overall consistency judgment, and whether the data passes the audit is determined
Updating the auditing state of the insurance transaction data on the basis of judging the consistency of the information. If the information is consistent, the data is confirmed to pass the audit, and the system state is updated. And updating the auditing state of the insurance transaction data into 'pass' by the system according to the consistency judging result. The updated state is recorded in the system database, ensuring subsequent tracking and processing. The system generates confirmation information of the passing of the audit, marks the insurance business form as the passing of the audit, and updates the processing progress of the insurance application. The status update information and related data are archived for subsequent auditing and querying.
Records of the passing audit are generated and stored. The record contains detailed information of the audit decision for subsequent use and audit. A detailed audit pass record report is generated recording updated information of the audit state, a summary of the comparison analysis results, and any relevant notes. The generated records are consolidated into a standardized format for storage and subsequent access. The audit passing records are stored in a database of an insurance company, so that the integrity and traceability of the data are ensured. And (5) backing up and archiving the audit record to ensure long-term storage and investigation.
And when the comparison and analysis result is that the information is consistent, the insurance transaction data is confirmed to pass the auditing, so that the automation level and the processing efficiency of the information auditing are improved. The automatic comparison process reduces manual intervention, reduces the risk of operation errors, and accelerates the auditing speed. The accurate information consistency judgment and state updating ensure the integrity and accuracy of the insurance transaction process, thereby optimizing the processing efficiency of insurance business, enhancing customer experience and improving the reliability and efficiency of the whole business process.
The invention relates to an intelligent auditing method for insurance data, which comprises the steps of acquiring customer verification data, wherein the data comprises an identity card picture and a target insurance certificate picture and is used for auditing insurance transaction data. And identifying the customer verification data by using the pre-trained identification model to generate a verification data identification result containing the target insurance certificate information. And extracting audit point information from the verification data identification result according to a predefined audit standard. And comparing and analyzing the audit point information with the corresponding information in the insurance transaction information to judge whether the information in the insurance transaction information is consistent with the information in the customer verification information. If the comparison and analysis result shows that the information is consistent, confirming that the insurance transaction data passes the audit. Through automatic processing and intelligent recognition, the invention improves auditing efficiency, reduces manual intervention and error rate, thereby accelerating the processing speed of insurance business orders and improving customer experience and business processing accuracy.
In one embodiment, after determining whether the information in the insurance transaction information is consistent with the information in the customer verification information in S40, the method further includes:
S401, if the comparison and analysis result is that the information is inconsistent, confirming that the insurance transaction data does not pass the audit;
S402, identifying and determining fields and corresponding positions in the insurance transaction information, wherein the fields and the positions are inconsistent with the audit point information;
S403, generating audit prompt information based on the inconsistent fields and the corresponding positions, wherein the audit prompt information comprises data items needing to be re-submitted or to be submitted in a supplementing mode;
s404, the audit prompt information is sent to the salesman end to prompt the resubmit or the supplement and submit of audit point information which does not pass the audit in the insurance transaction information.
In this embodiment, during the comparison process, if the information in the insurance transaction material is found to be inconsistent with the information in the customer verification material, the system will confirm that the insurance transaction material has not passed the audit. This process ensures that inconsistent information is handled correctly and measures are taken so that the customer can correct or supplement the material.
The system checks the results of the analysis and if any field is found inconsistent, marks the insurance transaction data as "audit failed". And updating the auditing state of the system to be 'failed', and recording inconsistent specific information. The system identifies from the comparison analysis which fields are inconsistent, such as insurance policy number, applicant name, etc. Marking the specific location of inconsistent fields in the insurance transaction information helps to accurately locate the problem. The contents of the inconsistent fields, the differences between expected and actual values, and the location of the fields in the material are recorded in detail.
And integrating all inconsistent fields and positions into a detailed record for generating audit prompt information. Based on the identified inconsistent fields and locations, detailed audit prompt information is generated. Such information includes items that need to be resubmitted or replenished. The prompt information is arranged into a standard format, so that the prompt information is clear and easy to understand. The inconsistent fields and positions are integrated into the prompt message to provide a clear indication. The generated reminder information will be used to inform the salesman or the customer so that he can take the corresponding action. The audit prompt is sent to the salesman through a suitable channel (e.g., email, text message or system notification). The notification comprises detailed audit prompt information which prompts a service person to resubmit or supplement audit point information which fails the audit. The system tracks the feedback of the prompt information by the service personnel, provides necessary support and guidance, and ensures that the problem is solved.
According to the embodiment, the auditing point information and the corresponding information in the insurance transaction information are compared and analyzed, so that the system can automatically detect and process the condition of inconsistent information, manual intervention is reduced, and data accuracy and processing efficiency are improved. The insurance business auditing process is optimized, the information processing speed is accelerated, meanwhile, the integrity and the reliability of the data are ensured, and finally, the customer experience and the business efficiency are improved.
In one embodiment, in S10, obtaining the client verification data includes:
s101, generating an H5 data uploading short link, wherein the uploading short link comprises a URL of an uploading page and parameters for identifying customer identity information and data types;
S102, sending the uploading short link to a mobile terminal of a client through a short message;
S103, when the uploading short link is triggered, providing an uploading page, and displaying corresponding data uploading options according to the data type;
s104, receiving the identity card picture and the target insurance certificate picture uploaded by the data uploading option as the customer verification data;
S105, checking whether the client verification data meets the format requirement;
S106, storing the customer verification data meeting the format requirements into a database, and associating the customer verification data with the customer identity information.
In this embodiment, an H5 short link is created that contains the upload page URL and parameters identifying the customer identity information and the profile type. A URL is generated for an upload page for receiving authentication material uploaded by a client. Parameters for identifying the identity of the client and the type of the required uploading data are embedded in the short link, so that the link can uniquely identify the client and the uploading requirement thereof. The use of short link generation tools shortens the complete URL and parameters to short links that are easy to send.
And sending the generated H5 short chain access short message service to the mobile terminal of the client. And sending the generated short chain to the mobile phone number of the client through the short message sending service platform. Ensure that the message contains a clear description informing the client of the type of data required for the link up-link.
When the upload short link is triggered, for example, after the client clicks on the short link, the system provides an upload page and displays the corresponding upload options according to the data type. The short link guides the client to access the uploading page, and the page displays corresponding uploading options according to the data types (such as identity cards and insurance certificates) which the client needs to upload. Relevant uploading fields and guides are displayed on the page, so that clients can conveniently and accurately upload data.
And receiving an identity card picture and a target insurance certificate picture submitted by the client through the uploading page. And the client submits the picture file of the identity card and the target insurance certificate through the uploading page. The system receives the uploaded pictures and performs preliminary verification, such as file format checking and size limiting.
And verifying whether the data uploaded by the client meets the preset format requirement. And checking whether the uploaded picture file meets the format (such as JPEG and PNG) required by the system or not, and whether the uploaded picture file meets the size and resolution requirements or not. For non-compliant materials, the system provides error prompts, requiring the client to re-upload the compliant files.
And storing the client verification data meeting the format requirements into a database and correlating with the client identity information. And storing the identity card and the insurance certificate picture file subjected to format verification in a database of the system. And (5) correlating the stored data with client identity information (such as client ID and mobile phone number) to ensure data integrity and traceability.
In the embodiment, the H5 data uploading short link is automatically generated and managed, so that the submitting process of the customer verification data is optimized. The automated upload page provides explicit upload guidelines, reducing errors in customer operations. The systematic file format check ensures the quality of data, and the automatic storage and information association improve the efficiency and accuracy of data management. The convenience of the customer verification process is improved, the requirement of manual operation is reduced, and the efficiency of overall business processing and customer experience are improved.
In one embodiment, in S20, before the identifying the customer verification material based on the pre-trained identification model, the method further includes:
S201, acquiring sample data, and dividing the sample data into a training set and a testing set;
S202, inputting the training set into a deep learning model for training, and adjusting the recognition parameters of the deep learning model root by learning the data characteristics in the training set to generate a preliminary recognition model;
S203, inputting the test set into the preliminary identification model to obtain a test result;
s204, analyzing the recognition accuracy of the preliminary recognition model based on the test result, and adjusting the recognition parameters of the preliminary recognition model according to the recognition accuracy;
And S205, training the preliminary recognition model for multiple times by using the test set until the recognition accuracy reaches an accuracy threshold, stopping training, and generating a final recognition model.
In this embodiment, sample data for training and testing is collected, including image data of customer verification material (e.g., identification card pictures, insurance certificate pictures, etc.), and divided into training sets and test sets. Representative sample data is obtained from various sources (e.g., historical data, public data sets, or manual collection), including image data under different conditions. The collected sample data is divided into training and testing sets at a certain ratio (e.g., 70% for training, 30% for testing). It is ensured that the training set and the test set contain different samples to improve the generalization ability of the model.
And inputting the divided training set data into a deep learning model for training so as to learn the characteristics in the image data and adjust the recognition parameters of the model. Training set data is input into the model using a deep learning framework (e.g., tensorFlow, pyTorch), model parameters (e.g., weights and biases) are adjusted through multiple iterations (epochs), optimizing the recognition capabilities of the model. Deep learning models progressively learn how to identify and classify data by extracting key features (e.g., text regions, characters, etc.) from images through Convolutional Neural Networks (CNNs) or other deep learning architecture.
And inputting the test set data into the trained preliminary recognition model, and obtaining a recognition result of the model for evaluating the performance of the model. The test set data is provided to a trained preliminary recognition model, which recognizes the test data and generates a prediction result (e.g., recognized text content, information fields, etc.). And carrying out preliminary analysis on the test result, and evaluating the performance of the model on unseen samples, wherein the performance indexes comprise the accuracy of identification, recall rate and the like.
And evaluating the performance of the preliminary recognition model by analyzing the accuracy of the test result, and adjusting the model parameters according to the recognition accuracy to optimize the recognition effect. The recognition accuracy of the model on the test set is calculated, including the ratio of correct recognition, the ratio of incorrect recognition, and the like. And adjusting super parameters (such as learning rate, regularization parameters and the like) of the model according to the analysis result, and optimizing the model to improve accuracy.
And training and adjusting the preliminary recognition model for multiple times until the recognition accuracy of the model reaches a preset accuracy threshold value, and finally generating an optimized recognition model. And repeating the training process, verifying the improvement condition of the model by using the test set, and updating parameters of the model after each training until the accuracy of the model on the test set reaches a preset threshold. And stopping training after the recognition accuracy of the model meets the requirement, and generating a final recognition model for practical application.
According to the embodiment, the client verification data is identified based on the pre-trained identification model, so that high accuracy and reliability of the identification model in the process of processing actual data can be ensured. Through multiple training and adjustment, the recognition performance of the model is optimized, the false recognition rate is reduced, and the accuracy of information extraction is improved. The finally generated identification model can effectively process verification data uploaded by clients, ensure the high efficiency and accuracy of the insurance handling process, and remarkably improve the speed and quality of business processing.
In one embodiment, the step S30 includes:
S301, defining auditing standards of the insurance transaction data based on target insurance business requirements;
S302, corresponding analysis is carried out according to the fields in the auditing standard and the fields in the identification result of the verification data, and auditing point information needing to be extracted is determined;
S303, extracting the audit point information, and formatting the audit point information based on the format standard of the insurance transaction data.
In this embodiment, in order to ensure the accuracy and integrity of the insurance transaction data, the audit standard is first defined according to the requirements of the target insurance business. These criteria will be used to extract key information from the verification material identification results. And (3) formulating a clear auditing standard to cover various information to be checked in the insurance transaction data. For example, for car insurance, the audit criteria may include owner name, insurance policy number, insurance amount, etc. Corresponding fields and data types are defined for each insurance type or business process, such as name on an identity card, an identity card number, an insurance policy number on an insurance certificate, an insurance policy amount, and the like.
And correspondingly analyzing the fields in the auditing standard and the fields identified from the verification data to determine which information needs to be extracted. And establishing a field mapping relation according to field definition in the auditing standard and field information in the verification data identification result. For example, the "PolicyNumber" field in the insurance credential identification result is mapped to the "policy number" field in the audit standard. The system identifies the fields meeting the auditing standards through comparison and analysis, and determines the information to be extracted.
And extracting key information fields meeting auditing standards from the identification result of the verification data. And the step is to sort the related information in the original identification result into audit point information. The system extracts field data matched with the auditing standard from the identification result. For example, the name and the identification card number are extracted from the identification card recognition result, and the insurance policy number and the insurance amount are extracted from the insurance credential recognition result. The extracted fields are arranged into a standard format for subsequent use and analysis.
Formatting the extracted audit point information to enable the audit point information to accord with the format standard of the insurance transaction data. And (5) formulating formatting standards according to requirements of insurance transaction information. For example, the identification number should be 15 or 18 digits and the policy number should conform to a particular character length and format. And (3) adjusting the extracted audit point information into a standardized format, and carrying out necessary conversion and correction to ensure that the format of the information is consistent with that of the insurance transaction information. And verifying whether the formatted data meets the standard or not, and carrying out necessary correction.
According to the embodiment, the auditing point information is extracted from the verification data identification result according to the predefined auditing standard, so that the accuracy and consistency of information extraction are ensured. The clear auditing standard and the automatic extraction process reduce the need of manual intervention, reduce the error rate and accelerate the data processing speed. The formatting process ensures the standard consistency of information and insurance handling data, optimizes the insurance auditing flow and improves the overall business efficiency and customer experience.
In one embodiment, the step S40 includes:
s405, extracting a data field matched with the type of the audit point information from the insurance transaction information as information to be audited;
s406, comparing each data field of the information to be checked with the corresponding field in the information of the checking point one by one;
S407, if each field in the information to be checked is completely consistent with the corresponding field in the information of the checking point, confirming that the comparison analysis result is information consistency, and if the data field inconsistent with the information of the checking point exists in the information to be checked, confirming that the comparison analysis result is information inconsistency;
S408, when the comparison analysis is completed, a comparison analysis result report is generated, and the comparison result of each field is recorded.
In this embodiment, a data field matching the audit point information type is extracted from the insurance transaction information. This step ensures that the data obtained from the insurance transaction is part of the comparison that is required. And identifying and extracting data fields matched with the information type of the audit point in the insurance transaction information. For example, the policy number and applicant name are extracted from the insurance application form and matched with corresponding fields extracted from the customer verification material. And arranging the extracted fields into a format of information to be checked, and preparing for further comparison.
And comparing each field of the information to be checked extracted from the insurance transaction information with the corresponding field in the information of the checking point item by item so as to confirm the consistency of the information. And comparing each field in the information to be audited with the corresponding field in the audit point information one by one. For example, the policy number is compared item by item with the policy number in the audit point information. The data format difference and possible errors are processed by comparing each field in detail using techniques such as character matching, data verification or regular expression.
And judging whether the information in the insurance transaction information is consistent with the information in the customer verification information or not based on the comparison result. And if each field in the information to be audited is completely consistent with the corresponding field in the audit point information, confirming that the information is consistent. The system updates the comparison result to be 'information consistent', and the insurance transaction data passes the auditing. And if any field to be checked is inconsistent with the corresponding field in the information of the checking point, recording the inconsistent condition of the data. The system updates the comparison result to 'information inconsistent', and the insurance transaction data does not pass the audit.
After the comparison analysis is completed, a report containing the comparison result is generated, and the comparison condition of each field is recorded. The comparison results, including the comparison results for each field, the consistency status, and any discovered inconsistent fields, are recorded in detail. The generated report is stored in a system database for subsequent review and processing.
According to the embodiment, the auditing point information and the corresponding information in the insurance transaction information are compared and analyzed, so that the automation level and accuracy of data auditing are improved. The systematic comparison process reduces manual operation, reduces error rate and accelerates processing speed. The accurate consistency judgment and the detailed comparison result report ensure the integrity and the reliability of the information, optimize the insurance auditing process and improve the overall business efficiency and the customer experience.
In one embodiment, in S50, after confirming that the insurance transaction information passes the audit, the method further includes:
s501, determining the service type of the insurance transaction data, and determining a service list template based on the service type;
S502, filling the information of the insurance transaction information into the service list template to generate an insurance service list;
S503, checking whether the information in the insurance business list is consistent with the corresponding audit point information;
S504, storing the verified and consistent insurance business list into an insurance archive.
In this embodiment, after confirming that the insurance transaction data passes the audit, the type of service transaction (e.g., car insurance, health insurance, etc.) needs to be determined first in order to select an appropriate service ticket template. The system automatically identifies the business type based on key information (e.g., insurance item name, product number, etc.) in the insurance transaction information. And classifying the insurance transaction data into corresponding business type categories according to predefined classification standards.
And selecting a proper template from a preset service list template library according to the identified service type, and generating an insurance service list. The system selects corresponding service list templates according to service types (such as automobile insurance templates and health insurance templates). The selected templates are applied to provide a format and layout basis for generating insurance policies.
And filling the information in the insurance transaction information into the selected service list template to generate a final insurance service list. Data in the insurance transaction material (such as the name of the applicant, insurance amount, insurance policy number and the like) is automatically filled into corresponding fields of the business form template. And generating a complete insurance business list, and ensuring that all relevant information is accurately displayed in the template.
And checking the generated insurance business ticket, and checking whether the information in the insurance business ticket is consistent with the information of the previous audit point. The system checks the information in the insurance business bill and compares the information with the information of the auditing points, so as to ensure that all the information is consistent. And checking each field by using a comparison tool to ensure that the data in the insurance business ticket is consistent with the audit point information.
And storing the verified and consistent insurance business list into an insurance archive for subsequent inquiry and management. And storing the verified insurance business list in an insurance archive to ensure data security and long-term storage. And an archive record is created for each insurance business bill, so that subsequent access, audit and data backtracking are facilitated.
According to the embodiment, after the insurance transaction data is checked, the system automatically selects the proper business form template and fills the information to generate the insurance business form, so that the accuracy and the processing efficiency of the generation of the insurance business form can be effectively improved. The automatic template selection and information filling process reduces manual intervention, reduces error rate and accelerates the generation speed of service orders. The verification process ensures that the generated business bill information is consistent with the audit point information, and further improves the accuracy of the data. Finally, the insurance business list with consistent verification is stored in an insurance archive, so that the integrity and traceability of data are ensured, the whole insurance handling process is optimized, and the business processing efficiency and customer satisfaction are improved.
The invention also provides an intelligent checking device for insurance data, referring to fig. 2, fig. 2 is a schematic diagram of functional modules of a preferred embodiment of the intelligent checking device for insurance data of the invention. The intelligent auditing device for insurance data comprises:
The data acquisition module is used for acquiring customer verification data, wherein the customer verification data comprises an identity card picture of a customer and a target insurance certificate picture, and is used for auditing insurance transaction data;
The data identification module is used for identifying the client verification data based on a pre-trained identification model and generating a verification data identification result, wherein the verification data identification result comprises target insurance certificate information;
the audit point extraction module is used for extracting audit point information from the verification data identification result according to a predefined audit standard;
The consistency auditing module is used for comparing and analyzing the auditing point information with the corresponding information in the insurance transaction information and judging whether the information in the insurance transaction information is consistent with the information in the customer verification information;
and the auditing result confirming module is used for confirming that the insurance transaction data passes auditing if the comparison and analysis results are consistent in information.
The specific implementation of the intelligent auditing device for insurance data is basically the same as the above embodiments of the intelligent auditing method for insurance data, and will not be repeated here.
The present invention also provides an insurance data intelligent auditing apparatus, as shown in fig. 3, which may include a processor 1001, such as a CPU, a communication bus 1002, a user interface 1003, a network interface 1004, and a memory 1005. Wherein the communication bus 1002 is used to enable connected communication between these components. The user interface 1003 may include a Display, an input unit such as a Keyboard (Keyboard), and the optional user interface 1003 may further include a standard wired interface, a wireless interface. The network interface 1004 may optionally include a standard wired interface, a wireless interface (e.g., WI-FI interface). The memory 1005 may be a high-speed RAM memory or a stable memory (non-volatile memory), such as a disk memory. The memory 1005 may also optionally be a storage device separate from the processor 1001 described above.
Those skilled in the art will appreciate that the hardware configuration of the insurance data intelligent auditing device shown in fig. 3 does not constitute a limitation of the insurance data intelligent auditing device, and may include more or fewer components than illustrated, or may combine certain components, or a different arrangement of components.
As shown in fig. 3, an operating system, a network communication module, a user interface module, and an insurance profile intelligent audit program may be included in the memory 1005 as one storage medium. The operating system is a program for managing and controlling the insurance data intelligent auditing equipment and software resources, and supports the operation of a network communication module, a user interface module, the insurance data intelligent auditing program and other programs or software, wherein the network communication module is used for managing and controlling a network interface 1004, and the user interface module is used for managing and controlling a user interface 1003.
In the hardware structure of the insurance data intelligent auditing apparatus shown in fig. 3, the network interface 1004 is mainly used for connecting a background server and performing data communication with the background server, the user interface 1003 is mainly used for connecting a client and performing data communication with the client, and the processor 1001 may call the insurance data intelligent auditing program stored in the memory 1005 and perform the same operation as the insurance data intelligent auditing method.
The specific implementation of the intelligent audit device for insurance data is basically the same as the above embodiments of the intelligent audit method for insurance data, and will not be described herein.
In addition, the embodiment of the invention also provides a computer storage medium, and the computer storage medium is stored with an insurance data intelligent auditing program, and the insurance data intelligent auditing program realizes the steps of the insurance data intelligent auditing method when being executed by a processor.
The specific implementation of the computer storage medium is basically the same as the above embodiments of the intelligent audit method of insurance data, and will not be described herein.
It should be noted that, if a software tool or component other than the company appears in the embodiment of the present application, the embodiment is merely presented by way of example, and does not represent actual use.