Data verification method, device, computer equipment and computer-readable storage medium

Abstract

The present application discloses a data verification method, apparatus, computer device, and computer-readable storage medium, which relate to the field of Internet technologies and can achieve a complete analysis of associated data, thereby improving the accuracy of the verification result of a target case. The method includes: obtaining a plurality of sample cases from a historical verification database, and determining a verification result corresponding to each sample case according to a plurality of sample associated data corresponding to each sample case; comparing the verification result with the review result corresponding to the sample case, and counting the number of correct samples with consistent verification result and review result content; when the ratio of the number of correct samples to the total number of samples of the plurality of sample cases is greater than or equal to a preset accuracy threshold, obtaining a verification rule model; obtaining a target case waiting for data verification from a case database, and inputting the target case into the verification rule model for data verification to obtain a target verification result of the target case.

Description

Technical Field

[0001] This application relates to the field of Internet technologies, and particularly to a data verification method, apparatus, computer device, and computer-readable storage medium. Background Art

[0002] With the continuous progress of Internet technologies and the gradual development of the insurance industry, the number of insurance policies has started to climb. In order to strictly control policy risks, insurance platforms adopt manual review to underwrite the policies generated by the insurance platforms.

[0003] In related technologies, after receiving a case to be underwritten, the staff for manual underwriting will query relevant information such as self-underwriting data and disease labels based on the customer information corresponding to the case, and give suggestions for underwriting conclusions on their own according to the relevant information.

[0004] In the process of implementing the present invention, the applicant found that the related technologies have at least the following problems:

[0005] With the increase in the volume of business, the number of cases that need to be manually reviewed is also increasing, the processing cycle will become longer and longer, and the professionalism of reviewers is also uneven, resulting in less analysis of the associated factors of the reported cases in manual review, low timeliness of issuing letters, and low accuracy of verification results. Summary of the Invention

[0006] In view of this, this application provides a data verification method, apparatus, computer device, and computer-readable storage medium, mainly aiming to solve the problem that the number of cases that need to be manually reviewed is increasing, the processing cycle will become longer and longer, and the professionalism of reviewers is also uneven, resulting in less analysis of the associated factors of the reported cases in manual review, low timeliness of issuing letters, and low accuracy of verification results.

[0007] According to the first aspect of this application, a data verification method is provided, and the method includes:

[0008] Obtain a plurality of sample cases from the historical verification database, and determine the verification result corresponding to each sample case according to a plurality of sample associated data corresponding to each sample case, where the sample associated data is the business data generated by the sample customer corresponding to the sample case on the platform;

[0009] Compare the verification result with the review result corresponding to the sample case, and count the number of correct samples with consistent verification result and review result content;

[0010] When the ratio of the number of correct samples to the total number of samples of the plurality of sample cases is greater than or equal to a preset accuracy threshold, obtain a verification rule model;

[0011] Obtain a target case waiting for data verification in the case database, input the target case into the verification rule model for data verification, and obtain the target verification result of the target case.

[0012] Optionally, determining the verification result of each sample case according to the multiple sample associated data corresponding to each sample case in the multiple sample cases includes:

[0013] Read the text type information of the sample case, extract the sample customer information and sample review data corresponding to the sample case from the text type information, where the sample customer information is used to indicate the sample customer corresponding to the sample case, and the sample review data is used to indicate the case type corresponding to the sample case;

[0014] Call the external system connection interface, identify the customer tags of all business data in the external business database, determine multiple specified customer tags whose tag content is consistent with the sample customer information, and use the multiple business data corresponding to the multiple specified customer tags as the multiple sample associated data;

[0015] Query multiple verification rule items corresponding to the sample review data in the verification rule database, perform data verification on the multiple sample associated data according to the multiple verification rule items, and obtain multiple sub-verification results;

[0016] Among the multiple sub-verification results, divide at least one sub-verification result with the same result content into the same result group to obtain multiple result groups;

[0017] Determine the total result proportion corresponding to each result group in the multiple result groups to obtain multiple total result proportions;

[0018] Sort the multiple total result proportions, determine a specified total result proportion whose ranking meets the preset conditions, and use the sub-verification result corresponding to the specified total result proportion as the verification result corresponding to the sample case;

[0019] For each sample case in the multiple sample cases, obtain the multiple sample associated data corresponding to each sample case, determine the verification result of each sample case, and obtain the multiple verification results.

[0020] Optionally, performing data verification on the multiple sample associated data according to the multiple verification rule items to obtain multiple sub-verification results includes:

[0021] For each sample associated data in the multiple sample associated data, determine the data category of the sample associated data, and query the first specified verification rule item in the multiple verification rule items whose item identifier is consistent with the data category;

[0022] Verify the sample associated data according to the verification index corresponding to the first specified verification rule item, and extract the verification result hit by the sample associated data as the verification result corresponding to the sample associated data;

[0023] Verify each sample associated data among the multiple sample associated data to obtain the multiple verification results.

[0024] Optionally, determining the total result proportion corresponding to each result group among the multiple result groups to obtain multiple total result proportions includes:

[0025] For each result group among the multiple result groups, identify at least one verification result included in the result group;

[0026] Determine the second specified verification rule item corresponding to each sub-verification result among the at least one sub-verification result, and use the proportion weight corresponding to the second specified verification rule item as the result proportion corresponding to each sub-verification result to obtain at least one result proportion corresponding to the at least one sub-verification result;

[0027] Add up the at least one result proportion to obtain the total result proportion corresponding to each result group;

[0028] Calculate the total result proportion corresponding to each result group among the multiple result groups to obtain the multiple total result proportions.

[0029] Optionally, comparing the multiple verification results with the review results corresponding to the multiple sample cases, and in the multiple verification results, counting the number of correct samples whose verification results are consistent with the review results includes:

[0030] Identify the review result corresponding to each sample case among the multiple sample cases, and compare the review result corresponding to each sample case with the verification result corresponding to each sample case;

[0031] Determine the target sample cases among the multiple sample cases whose verification results are consistent with the review results, and add a review label for indicating correct content to the target sample cases;

[0032] Count the number of target sample cases with a review label for indicating correct content added to obtain the number of correct samples.

[0033] Optionally, after obtaining the target case waiting for data verification in the case database, inputting the target case into the verification rule model for data verification to obtain the target verification result of the target case, the method further includes:

[0034] Determine the verification completion time point for obtaining the target verification result, and continuously count the time interval between the current time point and the verification completion time point;

[0035] When the time interval reaches the preset time interval, obtain a new target case from the case database again, input the new target case into the verification rule model for data verification, and obtain the target verification result of the new target case.

[0036] Optionally, the method further includes:

[0037] Send the verification result to the audit terminal for display, and obtain the audit result uploaded by the audit terminal;

[0038] When the audit result indicates changing the verification result, extract the audit result and the modification reason uploaded by the audit terminal, mark the target case information with the audit result, and store the marked target case information in the historical verification database;

[0039] Extract the target verification rule item and the modification index from the modification reason, and query the target verification index corresponding to the target verification rule item in the verification rule database;

[0040] Update the target verification index with the modification index, and store the updated target verification index in the verification rule database.

[0041] According to the second aspect of the present application, a data verification device is provided, and the device includes:

[0042] A determination module, configured to obtain a plurality of sample cases from the historical verification database, and determine the verification result corresponding to each sample case according to a plurality of sample associated data corresponding to each sample case, where the sample associated data is business data generated by the sample customer corresponding to the sample case on the platform;

[0043] A comparison module, configured to compare the verification result with the audit result corresponding to the sample case, and count the number of correct samples with consistent verification result and audit result content;

[0044] A calculation module, configured to obtain a verification rule model when the ratio of the number of correct samples to the total number of samples of the plurality of sample cases is greater than or equal to a preset accuracy threshold;

[0045] A verification module, configured to obtain a target case waiting for data verification from the case database, input the target case into the verification rule model for data verification, and obtain the target verification result of the target case.

[0046] Optionally, the determining module is configured to read the text - type information of the sample case, extract the sample customer information and sample review data corresponding to the sample case from the text - type information, where the sample customer information is used to indicate the sample customer corresponding to the sample case, and the sample review data is used to indicate the case type corresponding to the sample case; call an external - system connection interface to identify the customer tags of all business data in an external business database, determine multiple specified customer tags whose tag content is consistent with the sample customer information, and use the multiple business data corresponding to the multiple specified customer tags as the multiple sample - associated data; query multiple verification rule items corresponding to the sample review data in a verification - rule database, perform data verification on the multiple sample - associated data according to the multiple verification rule items, and obtain multiple sub - verification results; among the multiple sub - verification results, divide at least one sub - verification result with consistent result content into the same result group to obtain multiple result groups; determine the total result proportion corresponding to each result group in the multiple result groups to obtain multiple total result proportions; sort the multiple total result proportions, determine a specified total result proportion whose ranking meets a preset condition, and use the sub - verification result corresponding to the specified total result proportion as the verification result corresponding to the sample case; for each sample case among the multiple sample cases, obtain the multiple sample - associated data corresponding to each sample case, determine the verification result of each sample case, and obtain the multiple verification results.

[0047] Optionally, for each sample - associated data among the multiple sample - associated data, the determining module is configured to determine the data category of the sample - associated data, and query a first - specified verification rule item in the multiple verification rule items whose item identifier is consistent with the data category; perform verification on the sample - associated data according to the verification index corresponding to the first - specified verification rule item, and extract the verification result hit by the sample - associated data as the sub - verification result corresponding to the sample - associated data; perform verification on each sample - associated data among the multiple sample - associated data to obtain the multiple sub - verification results.

[0048] Optionally, for each result group among the multiple result groups, the determining module is configured to identify at least one sub - verification result included in the result group; determine a second - specified verification rule item corresponding to each sub - verification result among the at least one sub - verification result, and use the proportion weight corresponding to the second - specified verification rule item as the result proportion corresponding to each sub - verification result to obtain at least one result proportion corresponding to the at least one sub - verification result; add the at least one result proportion to obtain the total result proportion corresponding to each result group; calculate the total result proportion corresponding to each result group in the multiple result groups to obtain the multiple total result proportions.

[0049] Optionally, the comparison module is configured to identify the review result corresponding to each sample case among the multiple sample cases, compare the review result corresponding to each sample case with the verification result corresponding to each sample case; determine a target sample case with consistent verification result and review result among the multiple sample cases, add a review label indicating correct content to the target sample case; count the number of target sample cases with a review label indicating correct content added, to obtain the correct sample quantity.

[0050] Optionally, the apparatus further comprises:

[0051] A statistics module, configured to determine the verification completion time point when obtaining the target verification result, continuously count the time interval between the current time point and the verification completion time point;

[0052] The verification module is configured to, when the time interval reaches a preset time interval, re-obtain a new target case from the case database, input the new target case into the verification rule model for data verification, to obtain the target verification result of the new target case.

[0053] Optionally, the apparatus further comprises:

[0054] A display module, configured to send the verification result to an audit terminal for display, and obtain the audit result uploaded by the audit terminal;

[0055] A marking module, configured to, when the audit result indicates changing the verification result, extract the audit result and the modification reason uploaded by the audit terminal, mark the target case information with the audit result, and store the marked target case information into the historical verification database;

[0056] An extraction module, configured to extract a target verification rule item and a modification index from the modification reason, and query the target verification index corresponding to the target verification rule item in the verification rule database;

[0057] A storage module, configured to update the target verification index with the modification index, and store the updated target verification index into the verification rule database.

[0058] According to a third aspect of the present application, there is provided a computer device, comprising a memory and a processor, the memory stores a computer program, and when the processor executes the computer program, the steps of the method according to any one of the first aspects above are implemented.

[0059] According to the fourth aspect of the present application, there is provided a computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, the steps of the method described in any one of the above first aspects are implemented.

[0060] With the above technical solution, a data verification method, device, computer device and computer-readable storage medium provided by the present application first obtains a plurality of sample cases from a historical verification database, and determines the verification result corresponding to each sample case according to a plurality of sample associated data corresponding to each sample case. Subsequently, the verification result is compared with the review result corresponding to the sample case, and the number of correct samples with consistent content between the verification result and the review result is counted. When the ratio of the number of correct samples to the total number of samples of the plurality of sample cases is greater than or equal to a preset accuracy threshold, a verification rule model is obtained. Finally, a target case waiting for data verification is obtained from the case database, and the target case is input into the verification rule model for data verification to obtain the target verification result of the target case. By using the computing power of the computer to exhaust the associated data stored in the platform database for the target case, and by comparing the associated data with the corresponding indicators, a verification rule model is generated to achieve a complete analysis of the associated data, thereby improving the accuracy of the verification result of the target case.

[0061] The above description is only an overview of the technical solution of the present application. In order to be able to understand the technical means of the present application more clearly, it can be implemented according to the content of the specification. And in order to make the above and other purposes, features and advantages of the present application more obvious and understandable, the specific embodiments of the present application are hereinafter specifically exemplified. Brief Description of the Drawings

[0062] By reading the detailed description of the preferred embodiments below, various other advantages and benefits will become clear to those of ordinary skill in the art. The drawings are only for the purpose of showing the preferred embodiments and are not considered to be a limitation of the present application. And throughout the drawings, the same reference numerals are used to represent the same components. In the drawings:

[0063] Figure 1 It shows a schematic flow chart of a data verification method provided by an embodiment of the present application;

[0064] Figure 2 It shows a schematic flow chart of a data verification method provided by an embodiment of the present application;

[0065] Figure 3 It shows a schematic structural diagram of a data verification device provided by an embodiment of the present application;

[0066] Figure 4 It shows a schematic structural diagram of a device of a computer device provided by an embodiment of the present application. Detailed Embodiments

[0067] Exemplary embodiments of the present application will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present application are shown in the drawings, it should be understood that the present application can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided so that the present application can be more thoroughly understood and the scope of the present application can be completely conveyed to those skilled in the art.

[0068] An embodiment of the present application provides a data verification method. As Figure 1 shown, the method includes:

[0069] 101. Obtain a plurality of sample cases from the historical verification database, and determine the verification result corresponding to each sample case according to the plurality of sample associated data corresponding to each sample case, where the sample associated data is business data generated by the sample customer corresponding to the sample case on the platform.

[0070] 102. Compare the verification result with the audit result corresponding to the sample case, and count the number of correct samples with consistent verification result and audit result content.

[0071] 103. When the ratio of the number of correct samples to the total number of samples of the plurality of sample cases is greater than or equal to a preset accuracy threshold, obtain a verification rule model.

[0072] 104. Obtain a target case waiting for data verification from the case database, input the target case into the verification rule model for data verification, and obtain the target verification result of the target case.

[0073] The method provided by the embodiment of the present application first obtains a plurality of sample cases from the historical verification database, and determines the verification result corresponding to each sample case according to the plurality of sample associated data corresponding to each sample case. Subsequently, the verification result is compared with the audit result corresponding to the sample case, and the number of correct samples with consistent verification result and audit result content is counted. When the ratio of the number of correct samples to the total number of samples of the plurality of sample cases is greater than or equal to a preset accuracy threshold, a verification rule model is obtained. Finally, a target case waiting for data verification is obtained from the case database, and the target case is input into the verification rule model for data verification to obtain the target verification result of the target case. By using the computing power of the computer, the associated data stored in the platform database for the target case is exhausted, and by comparing the associated data with the corresponding metrics, a verification rule model is generated to achieve a complete analysis of the associated data, thereby improving the accuracy of the verification result of the target case.

[0074] An embodiment of the present application provides a data verification method. As Figure 2 shown, the method includes:

[0075] 201. Obtain multiple sample cases from the historical verification database, and determine the verification result corresponding to each sample case based on the multiple sample correlation data corresponding to each sample case.

[0076] With the continuous progress of Internet technology and the gradual development of the insurance industry, the number of insurance policies has begun to climb. In order to strictly control the policy risks, the insurance platform adopts the method of manual review to underwrite the policies generated by the insurance platform. At present, after receiving the cases that need to be underwritten, the staff of manual underwriting will query the relevant information such as self-verification data and disease labels according to the customer information corresponding to the cases, and give suggestions on underwriting conclusions by themselves according to the relevant information. However, the applicant realizes that with the increase in the business volume, the number of cases that need to be manually reviewed is increasing, the processing cycle will be longer and longer, and the professionalism of the reviewers is uneven, resulting in less analysis of the associated factors of the reported cases in manual review, low timeliness of letter issuance, and low accuracy of verification results.

[0077] Therefore, a data verification method, device, computer device and computer-readable storage medium provided by this application first obtain multiple sample cases from the historical verification database, and determine the verification result corresponding to each sample case based on the multiple sample correlation data corresponding to each sample case. Subsequently, compare the verification result with the review result corresponding to the sample case, and count the number of correct sample cases where the verification result is consistent with the review result content. When the ratio of the number of correct sample cases to the total number of sample cases of multiple sample cases is greater than or equal to the preset accuracy threshold, obtain the verification rule model. Finally, obtain the target case waiting for data verification from the case database, input the target case into the verification rule model for data verification, and obtain the target verification result of the target case. Utilize the computing power of the computer to exhaust the associated data stored in the platform database for the target case, generate a verification rule model by comparing the associated data with the corresponding indicators, realize the complete analysis of the associated data, and thus improve the accuracy of the target case verification result.

[0078] This application is applied to the human underwriting system, constructs a verification rule model based on the sample cases with known review results, and generates a target verification result by inputting the target sample case to be verified into the verification rule model. In addition, the human underwriting system can finally send the target verification result to the verification terminal for display, and continuously train the verification rule model according to the review results of the relevant staff on the target verification result, so as to obtain a verification rule model with higher accuracy.

[0079] In the embodiments of the present application, considering that platform staff will input the review results obtained through manual review into the historical verification database, a large amount of historical case information that has completed manual review is stored in the historical verification database. Specifically, the system first obtains multiple sample cases from the historical verification database, and determines multiple verification results corresponding to the multiple sample cases according to the multiple sample association data corresponding to each sample case in the multiple sample cases. For each sample case in the multiple sample cases, the specific process of obtaining the verification result is divided into the following steps:

[0080] Step 1: Identify each sample case and determine the multiple sample association data corresponding to the sample case.

[0081] In this step, for each sample case among multiple sample cases, the text - type information of the sample case can be read, and the sample customer information used to indicate the sample customer corresponding to the sample case and the sample review data used to indicate the case type corresponding to the sample case can be extracted from the text - type information. For example, customer number, customer name, customer gender, customer age, ID number, etc. are extracted as sample customer information, and policy type, reason for case initiation, etc. are extracted as sample review data. In the actual application process, the content categories of the extracted sample customer information and sample review data can adopt the system - default content categories or can be set by relevant staff according to the actual operation situation. The present invention does not specifically limit the types and contents included in the sample customer information and sample review data. Subsequently, an external system connection interface is called to identify the customer tags of all business data in the external business database. Among all the customer tags, multiple specified customer tags whose tag contents are consistent with the sample customer information are determined, and the multiple business data corresponding to the multiple specified customer tags are used as multiple sample - associated data. It should be noted that the sample - associated data is used to indicate all the business behavior data of the sample customer generated on the insurance platform. For example, self - review data, risk - control tags, preservation records, past disease records, claim records, etc. The present application does not specifically limit the types of the sample - associated data. In addition, since different types of associated data are stored in the underwriting databases corresponding to different systems of the platform, after the system obtains the sample customer information, it is necessary to call the external system link interface to connect with other systems of the platform. Then, according to the sample customer information, the sample - associated data corresponding to the sample customer information is queried in the underwriting databases corresponding to other systems. Specifically, the system queries the past manual underwriting information of the customer from the human underwriting system according to the sample customer information, that is, the insured information, obtains the past risk - control data from the risk - control intelligent underwriting system, obtains the past self - review data from the self - review system, queries the past policy - purchase records of the customer from the policy system, and obtains the past preservation and claim records of the customer from the preservation and claim systems. In the actual application process, considering that the databases corresponding to each system will add corresponding customer tags when storing behavior data, such as customer number tags, basic information tags, etc., so that the system can compare by retrieving the corresponding customer tags with the customer number stored in the sample customer information and extract the behavior data corresponding to the sample customer information.

[0082] By connecting with an external system and utilizing the computing power of the computer, all the sample - associated data of the sample customer is obtained, thus providing a complete data basis for generating a verification result based on the subsequent analysis of the sample - associated data.

[0083] Step 2: Based on the sample review data, query the validation rules corresponding to each sample case, and use the validation rules to validate multiple sample-related data to obtain the sub-validation results corresponding to each sample-related data.

[0084] In this step, the system first queries multiple validation rule items corresponding to the sample review data in the validation rule database, and performs data validation on the multiple sample-related data according to the multiple validation rule items to obtain multiple sub-validation results. It should be noted that the platform can upload the underwriting rules indicated by the text information to the system, and the system identifies the underwriting rules to determine the case types indicated in the underwriting rules, and adds labels to the underwriting rules using the case types.

[0085] In the actual application process, for each sample-related data among the multiple sample-related data, determine the data category of the sample-related data, such as self-underwriting data, risk control labels, preservation records, past disease records, claim records, etc. Query the first specified validation rule item in the multiple validation rule items whose item identifier is consistent with the data category. Subsequently, validate the sample-related data according to the validation indicators corresponding to the first specified validation rule item, and extract the validation results hit by the sample-related data as the sub-validation results corresponding to the sample-related data. For example, determine whether the self-underwriting is passed. If the self-underwriting is not passed, indicate that the sub-validation result corresponding to the sample-related data is "premium increase". Finally, validate each sample-related data among the multiple sample-related data to obtain multiple sub-validation results.

[0086] Step 3: Determine the validation result based on the multiple sub-validation results.

[0087] In fact, among multiple sub-verification results, at least one sub-verification result with consistent result content is divided into the same result group to obtain multiple result groups. For example, sub-verification results with the conclusion of "premium increase" are divided into the same result group, and sub-verification results with the conclusion of "exclusion" are divided into the same result group, etc. Subsequently, the total result proportion corresponding to each result group among the multiple result groups is determined to obtain multiple total result proportions. Specifically, for each result group among the multiple result groups, at least one verification result included in the result group is identified, and the second specified verification rule item corresponding to each verification result among the at least one verification result is determined. The proportion weight corresponding to the second specified verification rule item is used as the result proportion corresponding to each verification result to obtain at least one result proportion corresponding to the at least one verification result. The at least one result proportion is added up to obtain the total result proportion corresponding to each result group, and the total result proportion corresponding to each result group among the multiple result groups is calculated to obtain multiple total result proportions. Next, the multiple total result proportions are sorted, and the specified total result proportion whose ranking meets the preset conditions is determined, and the sub-verification result corresponding to the specified total result proportion is used as the verification result corresponding to the sample case. For example, the sample audit data sent by customer A to the system indicates that the policy type is a claim service type. Then, the underwriting rules corresponding to the claim service type are obtained. The self-underwriting is not passed, and "premium increase" is output, with a proportion weight of 20%. The risk control label shows that a major disease has occurred exceeding the claim threshold, and "premium increase" is output, with a proportion of 50%. The risk control label indicates false disclosure, and "rejection" is output, with a proportion of 10%. The preservation record indicates none, and normal claim payment is output with a proportion of 10%. The claim record indicates none, and "normal claim payment" is output with a proportion of 5%. The past medical history record indicates none, and "normal claim payment" is output with a proportion of 5%. Finally, the audit result "premium increase" is output according to the proportion ranking.

[0088] According to the sample customer information stored in the sample case, the previous associated data of the sample customer is found, including self-underwriting data, disease labels, previous preservation, previous claims and other associated information. Through model training, the verification result corresponding to the sample case is obtained, which greatly strengthens the analysis degree of the human underwriting system for the associated data, and thus improves the efficiency and accuracy of generating the verification result.

[0089] 202. Compare the verification result with the audit result corresponding to the sample case, and count the number of correct samples with consistent content between the verification result and the audit result.

[0090] In the embodiment of the present application, by comparing the verification result corresponding to each sample case with its corresponding audit result, it is judged whether the verification result generated by the verification rule model is consistent with the audit result. That is to say, if the content is consistent, it means that the verification result is correct; if the content is inconsistent, it means that the verification result is wrong.

[0091] Specifically, identify the review results corresponding to each sample case among multiple sample cases, and compare the review results corresponding to each sample case with the verification results corresponding to each sample case. Determine the target sample cases among the multiple sample cases where the verification results are consistent with the review results, and add a review label indicating correct content to the target sample cases. Further, count the number of target sample cases with a review label indicating correct content to obtain the correct sample quantity.

[0092] Through the above steps, add a review label indicating correct content to the sample cases where the verification results are consistent with the review results. Subsequently, by identifying the review labels, continuously count the number of sample cases with correct verification results, so as to calculate the accuracy rate of the verification rule model generating verification results based on the number of correct sample cases.

[0093] 203. When the ratio of the correct sample quantity to the total number of sample cases of the multiple sample cases is greater than or equal to the preset accuracy threshold, obtain the verification rule model.

[0094] In the embodiments of the present application, by calculating the ratio of the correct sample quantity to the total number of sample cases of the multiple sample cases, obtain the accuracy rate of the verification rule model generating verification results. When the accuracy rate is greater than or equal to the preset accuracy threshold, it indicates that the verification rule model can output correct verification results to a great extent. When the accuracy rate is less than the preset accuracy threshold, it is necessary to adjust the parameters of the verification rule model and continue to train the verification rule model.

[0095] Specifically, it is possible to identify the case type of the sample cases with verification errors, query the multiple verification rule items corresponding to this case type, and adjust the result proportion parameters corresponding to different verification rule items, thereby realizing the adjustment of the finally output verification results. Extract the information of multiple sample cases again and input it into the verification rule model to output the verification results, identify the case type of the sample cases with verification errors, and continuously adjust the model parameters of the verification rule model until the accuracy rate of generating verification results reaches the preset accuracy threshold.

[0096] Through the above steps, continuously extract multiple sample cases for verification, update the model parameters of the verification rule model, train the verification rule model, and finally complete the training of the verification rule model when the accuracy rate reaches the preset accuracy threshold to obtain the verification rule model.

[0097] 204. Obtain the target case waiting for data verification in the case database, input the target case into the verification rule model for data verification, and obtain the target verification result of the target case.

[0098] In an embodiment of the present application, the system obtains a target case to be verified from the case database, inputs the target case into a verification rule model for data verification, and outputs a target verification result of the target case, so that relevant staff can conduct case review based on the target verification result.

[0099] Among them, the case database contains information on all underwriting cases with successful reporting. Considering that there may be data content unrelated to verification in the target case information sent to the manual underwriting system, such as home address, phone number, etc. Therefore, the system needs to clean the data of the obtained target case and only retain the data content related to verification. Specifically, based on natural language processing technology, the target case information is identified, and the data indicating customer identity information is extracted as the target customer information, and the data indicating the case type is extracted as the target review data, such as the insured information, reporting type, policy information, etc. Further, actually, the system can set a time interval to determine the verification completion time point of the target verification result, and continuously count the time interval between the current time point and the verification completion time point. When the time interval reaches the preset time interval, a new target case is obtained again from the case database, and the new target case is input into the verification rule model for data verification to obtain the target verification result of the new target case. It should be noted that the system can set the target number of cases to be captured, capture the target number of target cases each time, verify these cases, start timing from the completion of the last target case, and when the time interval reaches the preset time interval, obtain the target number of target cases from the case database again.

[0100] In another implementation scenario, after the verification rule model outputs the verification result, the system will send the generated verification result to the review terminal for display, so that the staff of the review terminal can conduct manual review on the verification result. After the relevant staff complete the review, they need to upload the review result to the system. Subsequently, the system will identify the review result. When the review result indicates to change the verification result, the review result and the reason for modification uploaded by the review terminal are extracted, the target case information is marked with the review result, and the marked target case information is stored in the historical verification database. Further, the target verification rule item and the modification index are extracted from the reason for modification, the target verification index corresponding to the target verification rule item is queried in the verification rule database, the target verification index is updated with the modification index, and the updated target verification index is stored in the verification rule database. For example, for the underwriting cases where the reviewer has an 80% probability of automatically changing the underwriting conclusion of "reject insurance" to "extra premium" when there is a preservation record in the claim settlement, and it is indicated in the modification description that the preservation record is not sufficient to reject insurance, the system will automatically modify the result ratio of the preservation record in the corresponding claim settlement case in this scenario to obtain a more accurate verification result next time. In addition, when the review result indicates to confirm the verification result, the target case information is marked with the verification result, and the marked target case information is stored in the historical verification database.

[0101] In the actual application process, the system sets up a unified interface for manual review of reported cases, and sends the cases that meet the requirements for underwriting to the system by calling this interface. After the system stores the target case information in the case database, it generates a reminder message for successful reporting and sends the reminder message for successful reporting to the external underwriting system to prompt successful reporting.

[0102] The method provided by the embodiment of the present application first obtains multiple sample cases in the historical verification database, and determines the verification result corresponding to each sample case according to multiple sample associated data corresponding to each sample case. Subsequently, the verification result is compared with the review result corresponding to the sample case, and the number of correct sample cases with consistent content between the verification result and the review result is counted. When the ratio of the number of correct sample cases to the total number of sample cases of multiple sample cases is greater than or equal to the preset accuracy threshold, the verification rule model is obtained. Finally, the target case waiting for data verification is obtained in the case database, and the target case is input into the verification rule model for data verification to obtain the target verification result of the target case. By using the computing power of the computer to exhaust the associated data stored in the platform database for the target case, and comparing the associated data with the corresponding indicators, the verification rule model is generated, realizing the complete analysis of the associated data, and further improving the accuracy of the verification result of the target case.

[0103] Further, as Figure 1For the specific implementation of the method, an embodiment of the present application provides a data verification device, as Figure 3 shown. The device includes: a determination module 301, a comparison module 302, a calculation module 303, and a verification module 304.

[0104] The determination module 301 is configured to obtain a plurality of sample cases from a historical verification database, and determine a verification result corresponding to each sample case according to a plurality of sample association data corresponding to each sample case, where the sample association data is business data generated by a sample customer corresponding to the sample case on the platform;

[0105] The comparison module 302 is configured to compare the verification result with the review result corresponding to the sample case, and count the number of correct samples with consistent contents between the verification result and the review result;

[0106] The calculation module 303 is configured to obtain a verification rule model when the ratio of the number of correct samples to the total number of sample cases is greater than or equal to a preset accuracy threshold;

[0107] The verification module 304 is configured to obtain a target case waiting for data verification from a case database, input the target case into the verification rule model for data verification, and obtain a target verification result of the target case.

[0108] In a specific application scenario, the determination module 301 is configured to, for each sample case among the multiple sample cases, read the text - type information of the sample case, extract the sample customer information and sample review data corresponding to the sample case from the text - type information, where the sample customer information is used to indicate the sample customer corresponding to the sample case, and the sample review data is used to indicate the case type corresponding to the sample case; call an external - system connection interface, identify the customer tags of all business data in an external business database, determine multiple specified customer tags whose tag content is consistent with the sample customer information, and use the multiple business data corresponding to the multiple specified customer tags as the multiple sample - associated data; query multiple verification - rule items corresponding to the sample review data in a verification - rule database, perform data verification on the multiple sample - associated data according to the multiple verification - rule items, and obtain multiple verification results; among the multiple verification results, divide at least one verification result with consistent result content into the same result group to obtain multiple result groups; determine the total - result proportion corresponding to each result group among the multiple result groups to obtain multiple total - result proportions; sort the multiple total - result proportions, determine a specified total - result proportion whose ranking meets a preset condition, and use the verification result corresponding to the specified total - result proportion as the verification result corresponding to the sample case; for each sample case among the multiple sample cases, obtain the multiple sample - associated data corresponding to each sample case, determine the verification result of each sample case, and obtain the multiple verification results.

[0109] In a specific application scenario, the determination module 301 is configured to, for each sample - associated data among the multiple sample - associated data, determine the data category of the sample - associated data, and query a first - specified verification - rule item in the multiple verification - rule items whose item identifier is consistent with the data category; perform verification on the sample - associated data according to the verification index corresponding to the first - specified verification - rule item, and extract the verification result hit by the sample - associated data as the verification result corresponding to the sample - associated data; perform verification on each sample - associated data among the multiple sample - associated data to obtain the multiple verification results.

[0110] In a specific application scenario, the determination module 301 is configured to, for each result group among the multiple result groups, identify at least one verification result included in the result group; determine a second - specified verification - rule item corresponding to each verification result among the at least one verification result, and use the proportion weight corresponding to the second - specified verification - rule item as the result proportion corresponding to each verification result to obtain at least one result proportion corresponding to the at least one verification result; add the at least one result proportion to obtain the total - result proportion corresponding to each result group; calculate the total - result proportion corresponding to each result group among the multiple result groups to obtain the multiple total - result proportions.

[0111] In a specific application scenario, the comparison module 302 is configured to identify the review result corresponding to each sample case among the multiple sample cases, compare the review result corresponding to each sample case with the verification result corresponding to each sample case; determine target sample cases among the multiple sample cases where the verification result is consistent with the review result, add a review label indicating correct content to the target sample cases; and count the number of target sample cases with the review label indicating correct content to obtain the correct sample quantity.

[0112] In a specific application scenario, the device further includes: a statistics module 305.

[0113] The statistics module 305 is configured to determine the verification completion time point when the target verification result is obtained, and continuously count the time interval between the current time point and the verification completion time point;

[0114] The verification module 304 is configured to, when the time interval reaches a preset time interval, re-obtain a new target case from the case database, input the new target case into the verification rule model for data verification, and obtain the target verification result of the new target case.

[0115] In a specific application scenario, the device further includes: a display module 306, a marking module 307, an extraction module 308, and a storage module 309.

[0116] The display module 306 is configured to send the verification result to an audit terminal for display, and obtain the audit result uploaded by the audit terminal;

[0117] The marking module 307 is configured to, when the audit result indicates changing the verification result, extract the audit result and the modification reason uploaded by the audit terminal, mark the sample case information with the audit result, and store the marked sample case information in the historical verification database;

[0118] The extraction module 308 is configured to extract a target verification rule item and a modification index from the modification reason, and query the target verification index corresponding to the target verification rule item in the verification rule database;

[0119] The storage module 309 is configured to update the target verification index with the modification index, and store the updated target verification index in the verification rule database.

[0120] The device provided by the embodiment of the present application first obtains a plurality of sample cases from the historical verification database, and determines the verification result corresponding to each sample case according to a plurality of sample association data corresponding to each sample case. Subsequently, the verification result is compared with the review result corresponding to the sample case, and the number of correct samples with consistent verification result and review result content is counted. When the ratio of the number of correct samples to the total number of samples of the plurality of sample cases is greater than or equal to the preset accuracy threshold, a verification rule model is obtained. Finally, a target case waiting for data verification is obtained from the case database, and the target case is input into the verification rule model for data verification to obtain the target verification result of the target case. By using the computing power of the computer to exhaust the association data stored in the platform database for the target case, and by comparing the association data with the corresponding indicators, a verification rule model is generated to achieve a complete analysis of the association data, thereby improving the accuracy of the verification result of the target case.

[0121] It should be noted that for other corresponding descriptions of each functional unit involved in the data verification device provided by the embodiment of the present application, reference can be made to Figure 1 and Figure 2 the corresponding descriptions therein, which will not be elaborated here.

[0122] In an exemplary embodiment, referring to Figure 4 , a device is further provided. The device includes a bus, a processor, a memory, and a communication interface, and may further include an input / output interface and a display device. Among them, each functional unit can complete mutual communication through the bus. The memory stores a computer program, and the processor is used to execute the program stored on the memory to execute the data verification method in the above embodiment.

[0123] A computer-readable storage medium stores a computer program thereon, and when the computer program is executed by a processor, the steps of the data verification method are implemented.

[0124] Through the description of the above embodiments, those skilled in the art can clearly understand that the present application can be implemented through hardware or by means of software plus a necessary general hardware platform. Based on such an understanding, the technical solution of the present application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (which can be a CD-ROM, a USB flash drive, a mobile hard disk, etc.), including several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in various implementation scenarios of the present application.

[0125] Those skilled in the art can understand that the drawings are only schematic diagrams of a preferred implementation scenario, and the modules or processes in the drawings are not necessarily essential for implementing the present application.

[0126] Those skilled in the art can understand that the modules in the devices in the implementation scenarios can be distributed in the devices in the implementation scenarios according to the descriptions of the implementation scenarios, or can be correspondingly changed and located in one or more devices different from the present implementation scenario. The modules in the above implementation scenarios can be combined into one module, or can be further split into multiple sub-modules.

[0127] The above serial numbers of the present application are only for description and do not represent the advantages or disadvantages of the implementation scenarios.

[0128] The above discloses only several specific implementation scenarios of the present application. However, the present application is not limited thereto, and any changes that can be thought of by those skilled in the art should fall within the protection scope of the present application.

Claims

1. A data verification method, characterized in that, Including: Obtain multiple sample cases from the historical verification database, extract corresponding sample customer information and sample audit data from the text-type information of the sample cases, determine multiple sample association data corresponding to each sample case based on the sample customer information corresponding to each sample case, perform data verification on the multiple sample association data according to multiple verification rule items corresponding to the sample audit data, obtain multiple sub-verification results, divide at least one sub-verification result with consistent result content into the same result group, and determine the total result proportion corresponding to each result group. Take the sub-verification result corresponding to the specified total result proportion whose total result proportion ranking meets the preset condition as the verification result corresponding to the sample case. The sample association data is the business data generated by the sample customer corresponding to the sample case on the platform; Compare the verification result with the audit result corresponding to the sample case, and count the number of correct sample cases where the verification result is consistent with the audit result content; When the ratio of the number of correct sample cases to the total number of sample cases of the multiple sample cases is greater than or equal to the preset accuracy threshold, obtain the verification rule model; Obtain a target case waiting for data verification from the case database, input the target case into the verification rule model for data verification, and obtain the target verification result of the target case.

2. The method according to claim 1, wherein The determining the verification result of each sample case based on the multiple sample association data corresponding to each sample case in the multiple sample cases includes: Read the text-type information of the sample case, and extract the sample customer information and sample audit data corresponding to the sample case from the text-type information. The sample customer information is used to indicate the sample customer corresponding to the sample case, and the sample audit data is used to indicate the case type corresponding to the sample case; Call the external system connection interface, identify the customer tags of all business data in the external business database, determine multiple specified customer tags whose tag content is consistent with the sample customer information, and use the multiple business data corresponding to the multiple specified customer tags as the multiple sample association data; Query multiple verification rule items corresponding to the sample audit data in the verification rule database, and perform data verification on the multiple sample association data according to the multiple verification rule items to obtain multiple sub-verification results; Among the multiple sub-verification results, divide at least one sub-verification result with consistent result content into the same result group to obtain multiple result groups; Determine the total result proportion corresponding to each result group in the multiple result groups to obtain multiple total result proportions; Sort the multiple total result proportions, determine the specified total result proportion whose ranking meets the preset condition, and take the sub-verification result corresponding to the specified total result proportion as the verification result corresponding to the sample case; For each sample case in the multiple sample cases, obtain the multiple sample association data corresponding to each sample case, determine the verification result of each sample case, and obtain the multiple verification results.

3. The method according to claim 2, wherein Performing data verification on the multiple sample association data according to the multiple verification rule items to obtain multiple sub-verification results, including: For each sample association data among the multiple sample association data, determining the data category of the sample association data, and querying in the multiple verification rule items for the first specified verification rule item whose item identifier is consistent with the data category; Verifying the sample association data according to the verification index corresponding to the first specified verification rule item, and extracting the verification result hit by the sample association data as the sub-verification result corresponding to the sample association data; Verifying each sample association data among the multiple sample association data to obtain the multiple sub-verification results.

4. The method according to claim 2, wherein Determining the total result proportion corresponding to each result group among the multiple result groups to obtain multiple total result proportions, including: For each result group among the multiple result groups, identifying at least one sub-verification result included in the result group; Determining the second specified verification rule item corresponding to each sub-verification result among the at least one sub-verification result, and taking the proportion weight corresponding to the second specified verification rule item as the result proportion corresponding to each sub-verification result, to obtain at least one result proportion corresponding to the at least one sub-verification result; Adding up the at least one result proportion to obtain the total result proportion corresponding to each result group; Calculating the total result proportion corresponding to each result group among the multiple result groups to obtain the multiple total result proportions.

5. The method according to claim 1, wherein Comparing the multiple verification results with the review results corresponding to the multiple sample cases, and in the multiple verification results, counting the number of correct samples whose verification results are consistent with the review results, including: Identifying the review result corresponding to each sample case among the multiple sample cases, and comparing the review result corresponding to each sample case with the verification result corresponding to each sample case; Determining, among the multiple sample cases, the target sample cases whose verification results are consistent with the review results, and adding a review label for indicating correct content to the target sample cases; Counting the number of target sample cases with the review label for indicating correct content added to obtain the number of correct samples.

6. The method according to claim 1, characterized in that After obtaining the target case to be verified in the case database, inputting the target case into the verification rule model for data verification to obtain the target verification result of the target case, the method further includes: Determining the verification completion time point when the target verification result is obtained, and continuously counting the time interval between the current time point and the verification completion time point; When the time interval reaches a preset time interval, re-obtaining a new target case in the case database, and inputting the new target case into the verification rule model for data verification to obtain the target verification result of the new target case.

7. The method according to claim 1, wherein The method further includes: Sending the verification result to a review terminal for display, and obtaining the review result uploaded by the review terminal; When the audit result indicates a change in the verification result, extract the audit result and the reason for modification uploaded by the audit terminal, mark the target case information with the audit result, and store the marked target case information in the historical verification database; Extract the target verification rule item and the modification index from the reason for modification, and query the target verification index corresponding to the target verification rule item in the verification rule database; Update the target verification index with the modification index, and store the updated target verification index in the verification rule database.

8. A data verification device, characterized in that, It includes: A determination module, configured to obtain multiple sample cases from the historical verification database, extract corresponding sample customer information and sample audit data from the text type information of the sample cases, determine multiple sample association data corresponding to each sample case according to the sample customer information corresponding to each sample case, perform data verification on the multiple sample association data according to multiple verification rule items corresponding to the sample audit data, obtain multiple sub-verification results, divide at least one sub-verification result with consistent result content into the same result group, and determine the total result proportion corresponding to each result group, and use the sub-verification result corresponding to the specified total result proportion whose total result proportion ranking meets the preset condition as the verification result corresponding to the sample case, where the sample association data is the business data generated by the sample customer corresponding to the sample case on the platform; A comparison module, configured to compare the verification result with the audit result corresponding to the sample case, and count the number of correct samples with consistent content between the verification result and the audit result; A calculation module, configured to obtain a verification rule model when the ratio of the number of correct samples to the total number of samples of the multiple sample cases is greater than or equal to a preset accuracy threshold; A verification module, configured to obtain a target case waiting for data verification from the case database, input the target case into the verification rule model for data verification, and obtain the target verification result of the target case.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images