Claim settlement auditing method, claim settlement auditing device, computer device and storage medium

Abstract

The embodiment of the application provides a claim settlement auditing method, a claim settlement auditing device, a computer device and a storage medium, and belongs to the technical field of artificial intelligence. The method comprises the following steps: obtaining to-be-audited claim settlement information of a to-be-claim-settlement user, wherein the to-be-audited claim settlement information comprises claim settlement disease information and to-be-audited past disease information; obtaining a pre-trained past disease matching model; inputting the claim settlement disease information into the past disease matching model for data enhancement to obtain target disease information, wherein the target disease information comprises the claim settlement disease information and target past disease information, and the target past disease information is used for updating the training past disease matching model; performing past disease matching on the to-be-audited claim settlement information according to the past disease matching model to obtain cause-effect label information; and determining a claim settlement result of the to-be-claim-settlement user according to the cause-effect label information. The embodiment of the application can improve the efficiency and accuracy of claim settlement auditing on the relationship between past diseases.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of artificial intelligence, in particular to a claim settlement auditing method, a claim settlement auditing device, a computer device and a storage medium. BACKGROUND

[0002] The previous illness is one of the important reasons for insurance claim settlement business, and the previous illness refers to the disease or health abnormality of the insured person before the insurance, that is, the cause and effect relationship between the current claim case and the history disease of the insured person. At present, the rationality of the claim settlement application is mainly identified by manual auditing of the previous illness relationship of the current claim case. However, the operation efficiency of manual auditing is low, and the accuracy is not high. Therefore, how to improve the efficiency and accuracy of the claim settlement auditing of the previous illness relationship has become a technical problem to be solved. SUMMARY

[0003] The main purpose of the embodiments of the present application is to provide a claim settlement auditing method, a claim settlement auditing device, a computer device and a storage medium, which can improve the efficiency and accuracy of the claim settlement auditing of the previous illness relationship.

[0004] To achieve the above purpose, a claim settlement auditing method is provided in the first aspect of the embodiments of the present application, and the method comprises:

[0005] Obtaining the claim settlement information of a user to be audited, the claim settlement information comprising claim disease information and previous illness information to be audited;

[0006] Obtaining a pre-trained previous illness matching model;

[0007] Inputting the claim disease information into the previous illness matching model for data enhancement to obtain target disease information, the target disease information comprising the claim disease information and target previous illness information, and the target previous illness information being used to update the training of the previous illness matching model;

[0008] According to the previous illness matching model, the claim settlement information to be audited is matched to obtain cause and effect label information, and the cause and effect label information is used to represent the cause and effect relationship between the claim disease information and the previous illness information to be audited;

[0009] According to the cause and effect label information, the claim settlement result of the user to be audited is determined.

[0010] In some embodiments, the claim settlement result of the user to be audited is determined according to the cause and effect label information, comprising:

[0011] When the cause and effect label information represents that the claim disease information and the previous illness information to be audited have a cause and effect relationship, it is determined that the claim settlement result is to cancel the claim settlement.

[0012] When the causal tag information indicates that the claim disease information and the to-be-audited past illness information do not have a causal relationship, it is determined that the claim result is to agree to the claim.

[0013] In some embodiments, the pre-trained past illness matching model is trained by the following method:

[0014] A first sample set is constructed, the first sample set including a plurality of first sample pair information, the first sample pair information including sample disease information, first sample past illness information, and sample tag information, the sample tag information being used to represent a causal relationship between the sample disease information and the first sample past illness information;

[0015] An initial past illness matching model is constructed using a BERT algorithm, the initial past illness matching model including a data enhancement sub-model and a past illness matching sub-model;

[0016] The sample disease information is input into the initial past illness matching model, and the sample disease information is data-enhanced according to the data enhancement sub-model to obtain target sample pair information, the target sample pair information including the sample disease information, target sample past illness information, and the sample tag information;

[0017] A second sample set is constructed according to the first sample set and the target sample pair information, the second sample set including a plurality of second sample pair information, the second sample pair information being the first sample pair information or the target sample pair information, the second sample pair information including the sample disease information, second past illness information, and the sample tag information, the second past illness information being the first sample past illness information or the target sample past illness information;

[0018] The sample disease information and the second past illness information are matched according to the past illness matching sub-model to obtain target tag information;

[0019] A tag loss value is determined according to the sample tag information and the target tag information;

[0020] Model parameters of the initial past illness matching model are adjusted according to the tag loss value, and the adjusted initial past illness matching model is continuously trained based on the second sample set until the tag loss value meets a preset training end condition, so as to obtain the past illness matching model.

[0021] In some embodiments, before the sample disease information is input into the initial past illness matching model and the sample disease information is data-enhanced according to the data enhancement sub-model to obtain target sample pair information, the method further includes:

[0022] adding a text mark at a text start part and a text end part of the sample disease information respectively, to update the sample disease information, the text mark including a text start mark corresponding to the text start part and a text end mark corresponding to the text end part, the text start mark being used to represent that the sample disease information starts to be processed, and the text end mark being used to represent that the sample disease information ends to be processed.

[0023] In some embodiments, the inputting the sample disease information into the initial past illness matching model, and performing data enhancement on the sample disease information according to the data enhancement sub-model to obtain target sample pair information, comprises:

[0024] inputting the sample disease information into the initial past illness matching model, and performing prediction processing on the sample disease information when the data enhancement sub-model detects the text start mark of the sample disease information to obtain first prediction text information;

[0025] adding the first prediction text information behind the text end mark of the sample disease information to obtain candidate sample information;

[0026] when it is detected that the first prediction text information is not the text end mark, inputting the sample disease information and the first prediction text information into the initial past illness matching model, and performing prediction processing on the sample disease information to obtain second prediction text information;

[0027] adding the second prediction text information behind the candidate sample information to update the candidate sample information;

[0028] setting the second prediction text information as the first prediction text information, and repeating the prediction processing according to the sample disease information and the first prediction text information until it is detected that the first prediction text information is the text end mark, and adding the text end mark behind the candidate sample information to obtain target sample pair information.

[0029] In some embodiments, the model parameters of the initial past illness matching model are used to train the data enhancement sub-model and the past illness matching sub-model, the model parameters of the initial past illness matching model are adjusted according to the label loss value, and the adjusted initial past illness matching model is continuously trained based on the second sample set until the label loss value meets a preset training end condition to obtain the past illness matching model, comprising:

[0030] adjusting the model parameters of the past illness matching sub-model according to the label loss value;

[0031] training the data augmentation sub-model according to the adjusted model parameters to update the second sample set;

[0032] continuing to train the adjusted initial previous illness matching model based on the second sample set until the label loss value meets a preset training end condition, to obtain the previous illness matching model.

[0033] To achieve the above object, a second aspect of the embodiment of the present application provides a claim settlement auditing device, which comprises:

[0034] an information acquisition module, configured to acquire to-be-audited claim settlement information of a to-be-claim-settlement user, the to-be-audited claim settlement information comprising claim settlement disease information and to-be-audited previous illness information;

[0035] a model acquisition module, configured to acquire a pre-trained previous illness matching model;

[0036] a data augmentation module, configured to input the claim settlement disease information into the previous illness matching model for data augmentation, to obtain target disease information, the target disease information comprising the claim settlement disease information and target previous illness information, the target previous illness information being used to update and train the previous illness matching model;

[0037] a previous illness matching module, configured to perform previous illness matching on the to-be-audited claim settlement information according to the previous illness matching model, to obtain cause-effect label information, the cause-effect label information being used to represent a cause-effect relationship between the claim settlement disease information and the to-be-audited previous illness information;

[0038] a claim settlement determination module, configured to determine a claim settlement result of the to-be-claim-settlement user according to the cause-effect label information.

[0039] To achieve the above object, a third aspect of the embodiment of the present application provides a computer device, which comprises a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to perform the method according to any one of the embodiments of the first aspect of the present application.

[0040] To achieve the above object, a fourth aspect of the embodiment of the present application provides a storage medium, which is a computer readable storage medium, and stores a computer program, and the computer program is executed by a computer to perform the method according to any one of the embodiments of the first aspect of the present application.

[0041] The claim settlement auditing method, the claim settlement auditing device, the computer device and the storage medium provided by the embodiments of the present application obtain the to-be-audited claim settlement information of a to-be-claim-settlement user, the to-be-audited claim settlement information including claim settlement disease information and to-be-audited past illness information. In order to solve the problem that the accuracy of claim settlement auditing of past illness relationship is low due to different writing methods of to-be-audited claim settlement information, a pre-trained past illness matching model is obtained, the claim settlement disease information is input into the past illness matching model for data enhancement, target disease information is obtained, the target disease information including the claim settlement disease information and target past illness information, and the target past illness information is used to update the training past illness matching model. Then, the to-be-audited claim settlement information is matched according to the past illness matching model, and cause-effect label information is obtained, the cause-effect label information being used to represent the cause-effect relationship between the claim settlement disease information and the to-be-audited past illness information. Finally, the claim settlement result of the to-be-claim-settlement user is determined according to the cause-effect label information. The present application introduces data enhancement in the model, enriches the data samples in the training stage, thereby improving the prediction effect of untrained samples, and further improving the efficiency and accuracy of claim settlement auditing of past illness relationship. BRIEF DESCRIPTION OF DRAWINGS

[0042] Figure 1 is a flowchart of the claim settlement auditing method provided by the embodiments of the present application;

[0043] Figure 2 is a training flowchart of the past illness matching model provided by the embodiments of the present application;

[0044] Figure 3 is a flowchart of step S230 in Figure 2

[0045] Figure 4 is a flowchart of step S270 in Figure 2

[0046] Figure 5 is a module structure block diagram of the claim settlement auditing device provided by the embodiments of the present application;

[0047] Figure 6 is a hardware structure schematic diagram of the computer device provided by the embodiments of the present application. DETAILED DESCRIPTION

[0048] In order to make the objectives, technical solutions and advantages of the present application clearer, the present application is further described in detail below with reference to the drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application, and are not used to limit the present application.

[0049] ​​It should be noted that, although the functional modules are divided in the device schematic diagram, and the logical order is shown in the flowchart, in some cases, the steps shown or described can be performed in a manner different from the module division in the device or the order in the flowchart. The terms "first", "second", and the like in the description and claims and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0050] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this application belongs. The terminology used in the specification herein is for describing the embodiments of the present application only and is not intended to limit the present application.

[0051] In addition, the described features, structures or characteristics can be combined in any suitable manner in one or more embodiments. In the following description, numerous specific details are provided to give a thorough understanding of embodiments of the present disclosure. One skilled in the relevant art will recognize, however, that the technical solutions of the present disclosure can be practiced without one or more of the specific details, or with other methods, components, devices, steps, etc. In other cases, well-known methods, devices, implementations or operations are not shown or described in detail to avoid obscuring aspects of the present disclosure.

[0052] The block diagrams shown in the accompanying drawings are only functional entities, and do not necessarily correspond to physically independent entities. That is, these functional entities can be implemented in the form of software, or in one or more hardware modules or integrated circuits, or in different networks and / or processor devices and / or microcontroller devices.

[0053] The flowcharts shown in the accompanying drawings are only exemplary illustrations, and do not necessarily include all contents and operations / steps, nor do they necessarily have to be executed in the order described. For example, some operations / steps can be further divided, and some operations / steps can be combined or partially combined, so the actual execution order may be changed according to the actual situation.

[0054] First, some terms involved in the present application are analyzed:

[0055] Artificial Intelligence (AI): is a new technical science that studies, develops theories, methods, technologies and application systems for simulating, extending and expanding human intelligence; Artificial intelligence is a branch of computer science, artificial intelligence attempts to understand the essence of intelligence and produce a new intelligent machine that can react in a similar way to human intelligence. The research in this field includes robots, language recognition, image recognition, natural language processing and expert systems, etc. Artificial intelligence can simulate the information process of human consciousness and thinking. Artificial intelligence is also the theory, method, technology and application system of using digital computer or digital computer controlled machine to simulate, extend and expand human intelligence, perceive environment, acquire knowledge and use knowledge to obtain the best results.

[0056] Data augmentation: refers to generating more data from limited data to increase the number of training samples and diversity (noise data), so as to improve the robustness of the model.

[0057] BERT (Bidirectional Encoder Representation from Transformers): used to further increase the generalization ability of word vector model, fully describe the character level, word level, sentence level and even inter-sentence relationship features, which is based on Transformer. There are three kinds of Embedding in BERT, namely Token Embedding, SegmentEmbedding and Position Embedding; among them, Token Embeddings is word vector, the first word is CLS mark, which can be used for subsequent classification tasks; Segment Embeddings is used to distinguish two sentences, because pre-training not only does LM but also does classification task with two sentences as input; Position Embeddings, the position word vector here is not the triangular function in transfor, but the position information learned by BERT through training. But BERT directly trains a PositionEmbedding to retain position information, initializes a vector for each position at random, adds it to the model training, and finally gets an Embedding containing position information. Finally, the combination of Position Embedding and word Embedding, BERT chooses to directly concatenate.

[0058] Attention Mechanism: Attention mechanism can enable neural networks to focus on a subset of their inputs (or features), selecting specific inputs, which can be applied to any type of input regardless of its shape. In the case of limited computing power, attention mechanism is a resource allocation scheme that allocates computing resources to more important tasks to solve the problem of information overload.

[0059] Pre-existing disease is one of the important reasons for insurance claim business to refuse compensation. Pre-existing disease refers to the disease or health abnormality that the insured has occurred on the body before the insurance is applied for, that is, there is a causal relationship between the current claim case and the history disease of the insured. At present, for the rationality judgment of claim application, artificial audit is mainly used to identify the pre-existing disease relationship of the current claim case. However, the operation efficiency of artificial audit is low, and the accuracy is not high. Therefore, how to improve the efficiency and accuracy of the claim audit of the pre-existing disease relationship has become a technical problem to be solved.

[0060] Based on this, the embodiment of the application provides a claim auditing method, a claim auditing device, a computer equipment and a storage medium, which can improve the efficiency and accuracy of the claim audit of the pre-existing disease relationship.

[0061] The claim auditing method, claim auditing device, computer equipment and storage medium provided by the embodiment of the application are specifically explained by the following embodiment. First, the claim auditing method in the embodiment of the application is described.

[0062] The embodiment of the application can acquire and process related data based on artificial intelligence technology. Artificial intelligence (AI) is the use of digital computers or computer-controlled machines to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0063] The basic technology of artificial intelligence generally includes technologies such as sensors, special artificial intelligence chips, cloud computing, distributed storage, big data processing technology, operation / interaction system, mechatronics, etc. The software technology of artificial intelligence mainly includes computer vision technology, robot technology, biometric technology, speech processing technology, natural language processing technology and machine learning / deep learning, etc.

[0064] The claim settlement auditing method provided by the embodiments of the present application relates to the field of artificial intelligence. The claim settlement auditing method provided by the embodiments of the present application can be applied to a terminal, can be applied to a server end, and can also be software running in the terminal or the server end. In some embodiments, the terminal can be a smart phone, a tablet computer, a notebook computer, a desktop computer, a smart watch or the like; the server end can be configured as a stand-alone physical server, can be configured as a server cluster or a distributed system formed by multiple physical servers, can also be configured as a cloud server providing basic cloud computing services such as cloud service, cloud database, cloud computing, cloud function, cloud storage, network service, cloud communication, middleware service, domain name service, security service, content delivery network (CDN) and big data and artificial intelligence platform; and the software can be an application or the like for implementing the claim settlement auditing method, but is not limited to the above forms.

[0065] The embodiments of the present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, and the like. The present application can be described in the general context of computer-executable instructions executed by a computer, such as a program module. Generally, program modules include routines, programs, objects, components, data structures, and the like that perform specific tasks or implement specific abstract data types. The present application can also be practiced in a distributed computing environment in which tasks are performed by remote processing devices connected by a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media, including storage devices.

[0066] Please refer to Figure 1 , Figure 1 is an optional flowchart of the claim settlement auditing method provided by the embodiments of the present application. In some embodiments, the claim settlement auditing method of the embodiments of the present application includes but is not limited to steps S110 to S150, which will be described below in combination with Figure 1 The five steps will be described in detail.

[0067] Step S110, obtaining claim settlement information to be audited of a user to be settled, the claim settlement information to be audited including claim settlement disease information and to-be-audited past disease information;

[0068] Step S120, obtaining a pre-trained past disease matching model;

[0069] In step S130, the claim disease information is input into the previous disease matching model for data enhancement, to obtain target disease information, the target disease information including the claim disease information and target previous disease information, the target previous disease information being used for updating the trained previous disease matching model.

[0070] In step S140, the previous disease matching model is used to match the previous disease of the to-be-audited claim information, to obtain cause-effect label information, the cause-effect label information being used to represent the cause-effect relationship between the claim disease information and the to-be-audited previous disease information.

[0071] In step S150, the claim result of the to-be-claimed user is determined according to the cause-effect label information.

[0072] In step S110 of some embodiments, in order to audit the previous disease relationship of the current claim case, to-be-claimed user to-be-audited claim information is obtained, the to-be-audited claim information including claim disease information and to-be-audited previous disease information. Specifically, the claim disease information is used to represent the disease complaint of the to-be-claimed user in the current claim case. The to-be-audited previous disease information is used to represent the previous disease matching information in the historical claim information of the to-be-claimed user, which may be related to the audit of the previous disease relationship.

[0073] It should be noted that the to-be-claimed user can include at least one claim disease information, that is, at least one disease complaint. In order to improve the efficiency of claim auditing, the information corresponding to the disease complaint can be selected as the claim disease information for claim auditing in the embodiments of the present application, and the disease complaint refers to the first disease complaint in the claim application file corresponding to the to-be-claimed user. Different disease complaints can also be determined according to requirements, so as to determine the to-be-audited claim information.

[0074] It should be noted that before step S110, the method of the embodiments of the present application can further include obtaining a previous disease matching request, the previous disease matching request including information of the to-be-claimed user and the to-be-audited claim information.

[0075] It should be noted that the historical claim information can obtain relevant historical claim data from the cloud server of the insurance company, and obtain the to-be-audited previous disease information according to keyword extraction and other text information matching.

[0076] It should be noted that in some embodiments, the information related to the previous disease matching in the historical claim information or the claim application file can also be obtained by writing a web crawler or a script program for targeted data crawling.

[0077] In step S120 of some embodiments, due to the fact that different doctors have different writing habits in matching the required fields of the past medical history during the historical diagnosis process, the disease name written varies greatly, resulting in a low accuracy of the claim settlement review of the past medical history relationship. In some embodiments, a pre-trained past medical history matching model is obtained. Specifically, the generation process of the sample is introduced into the pre-trained past medical history matching model, and the model has good generalization through data enhancement, thereby improving the accuracy of the claim settlement review of the past medical history relationship.

[0078] In step S130 of some embodiments, in order to improve the efficiency of the model in claim settlement review of the past medical history relationship, the claim settlement disease information is input into the past medical history matching model for data enhancement to obtain target disease information. The target disease information includes claim settlement disease information and target past medical history information, and the target past medical history information is used to update the training past medical history matching model. Specifically, in the actual application process, the claim settlement disease information can be input into the past medical history matching model for data enhancement to obtain the target disease information. The target disease information can be added to the training set corresponding to the model, and the model can be continuously trained according to the updated training set, thereby improving the efficiency of the model in claim settlement review of the past medical history relationship.

[0079] In step S140 of some embodiments, in order to accurately review and determine the past medical history relationship between the claim settlement disease information and the past medical history information to be reviewed in the current claim settlement case, the past medical history matching model is used to match the past medical history of the claim settlement information to be reviewed, and the causal label information is obtained. The causal label information is used to represent the causal relationship between the claim settlement disease information and the past medical history information to be reviewed.

[0080] In step S150 of some embodiments, the claim settlement result of the user to be settled is determined according to the causal label information. Specifically, according to the causal label information, the past medical history relationship between the claim settlement disease information and the past medical history information to be reviewed in the current claim settlement case can be determined, i.e., the corresponding claim settlement result is determined.

[0081] It should be noted that in each of the specific embodiments of the present application, when it is necessary to process relevant data related to the identity or characteristics of the user according to the user information, user behavior data, original claim settlement samples, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of these data will comply with relevant laws, regulations, and standards of the country or region. In addition, when the embodiments of the present application need to obtain sensitive personal information of the user, the separate permission or separate consent of the user will be obtained through a pop-up window or a jump to a confirmation page, and after obtaining the separate permission or separate consent of the user, the necessary user-related data for the normal operation of the embodiments of the present application will be obtained.

[0082] In some embodiments of the present application, step S150 specifically includes any one of the following:

[0083] When the causal tag information indicates that the claim disease information and the to-be-reviewed past medical history information have a causal relationship, it is determined that the claim result is to cancel the claim;

[0084] When the causal tag information indicates that the claim disease information and the to-be-reviewed past medical history information do not have a causal relationship, it is determined that the claim result is to approve the claim.

[0085] It should be noted that the causal tag information can be set to two, which are respectively labeled 0 and 1, and label 0 indicates a non-causal relationship and label 1 indicates a causal relationship. Specifically, when the causal tag information is label 1, it indicates that the claim disease information and the to-be-reviewed past medical history information have a causal relationship, and it is determined that the claim result is to cancel the claim; when the causal tag information is label 0, it indicates that the claim disease information and the to-be-reviewed past medical history information do not have a causal relationship, and it is determined that the claim result is to approve the claim, thereby realizing automatic review of the claim result and improving the efficiency and accuracy of claim review.

[0086] Please refer to Figure 2 , Figure 2 is a flowchart of a training process of a pre-trained past medical history matching model provided by an embodiment of the present application. In some embodiments, the training process of the past medical history matching model specifically includes but is not limited to steps S210 to S270, which will be described below in combination with Figure 2 The seven steps will be described in detail.

[0087] Step S210, a first sample set is constructed, the first sample set includes a plurality of first sample pair information, the first sample pair information includes sample disease information, first sample past medical history information and sample label information, the sample label information is used to represent the causal relationship between the sample disease information and the first sample past medical history information;

[0088] Step S220, an initial past medical history matching model is constructed using a BERT algorithm, the initial past medical history matching model includes a data enhancement sub-model and a past medical history matching sub-model;

[0089] Step S230, the sample disease information is input into the initial past medical history matching model, the data enhancement sub-model is used to perform data enhancement on the sample disease information, and target sample pair information is obtained, the target sample pair information includes sample disease information, target sample past medical history information and sample label information;

[0090] In step S240, a second sample set is constructed according to the first sample set and the target sample pair information, the second sample set includes a plurality of second sample pair information, the second sample pair information is the first sample pair information or the target sample pair information, and the second sample pair information includes sample disease information, second past medical history information, and sample label information, the second past medical history information is the first sample past medical history information or the target sample past medical history information.

[0091] In step S250, the past medical history matching is performed on the sample disease information and the second past medical history information according to the past medical history matching sub-model, and target label information is obtained.

[0092] In step S260, a label loss value is determined according to the sample label information and the target label information.

[0093] In step S270, the model parameters of the initial past medical history matching model are adjusted according to the label loss value, and the adjusted initial past medical history matching model is continuously trained based on the second sample set until the label loss value meets a preset training end condition, so as to obtain the past medical history matching model.

[0094] In step S210 of some embodiments, in order to train the past medical history matching model, first, a first sample set is constructed, the first sample set includes a plurality of first sample pair information, the first sample pair information includes sample disease information, first sample past medical history information, and sample label information, and the sample label information is used to represent the causal relationship between the sample disease information and the first sample past medical history information.

[0095] It should be noted that the setting of the sample label information is the same as the setting of the causal label information in the above-mentioned embodiments, and will not be repeated here. The acquisition of the sample disease information in the first sample pair information can refer to the acquisition of the claim disease information in the above-mentioned embodiments, and the acquisition of the first sample past medical history information can refer to the acquisition of the to-be-audited past medical history information in the above-mentioned embodiments, which will not be repeated here.

[0096] In step S220 of some embodiments, in order to improve the efficiency of the past medical history matching model in claim auditing of the past medical history relationship, an initial past medical history matching model is constructed by using a BERT algorithm, the initial past medical history matching model includes a data enhancement sub-model and a past medical history matching sub-model, that is, the training of data enhancement of the model and the training of past medical history matching of the model adopt the same model structure, so that the sharing of the model parameters is realized, and the efficiency of the past medical history matching model in claim auditing of the past medical history relationship is improved.

[0097] In step S230 of some embodiments, in order to improve the generalization ability of the previous illness matching model for processing of unobserved samples, the sample disease information is input into the initial previous illness matching model, the sample disease information is data-augmented according to the data augmentation sub-model, and target sample pair information is obtained, the target sample pair information including the sample disease information, target sample previous illness information, and sample label information. Specifically, the sample disease information is data-augmented according to the data augmentation sub-model, and new target sample pair information corresponding to the sample disease information is generated, the sample label information in the target sample pair information being the same as the sample label information in the first sample pair information. For example, when the first sample pair information includes sample disease information A, first sample previous illness information B, and sample label information C, the generated target sample pair information includes sample disease information A, target sample previous illness information B', and sample label information C.

[0098] In step S240 of some embodiments, in order to improve the efficiency of claim settlement review on the previous illness relationship, after obtaining the target sample pair information through data augmentation, a second sample set is constructed according to the first sample set and the target sample pair information, the second sample set including a plurality of second sample pair information, the second sample pair information being the first sample pair information or the target sample pair information, the second sample pair information including sample disease information, second previous illness information, and sample label information, the second previous illness information being the first sample previous illness information or the target sample previous illness information. Specifically, the second sample set is constructed according to the obtained target sample pair information and the first sample set, that is, the number of the sample set used for model training is expanded, and the augmentation of the sample set can effectively reduce the overfitting phenomenon of the model network and improve the generalization ability of the trained model network, better adapting to different application scenarios.

[0099] In steps S250 to S270 of some embodiments, the previous illness matching sub-model is used to match the sample disease information and the second previous illness information, and target label information is obtained. Specifically, the initial previous illness matching model is trained according to the obtained second sample set and the previous illness matching sub-model, the previous illness matching sub-model is used to match the sample disease information and the second previous illness information, and target label information is obtained, the target label information being the label information predicted by the previous illness matching sub-model. A label loss value is determined according to the sample label information and the target label information. Then, the model parameters of the initial previous illness matching model are adjusted according to the label loss value, and the adjusted initial previous illness matching model is continuously trained based on the second sample set until the label loss value meets a preset training end condition, so as to obtain the previous illness matching model.

[0100] It should be noted that the preset training end condition can be that the label loss value is less than a preset loss value threshold, or that the label accuracy of the initial previous illness matching model is greater than or equal to a preset accuracy threshold.

[0101] In some embodiments of the present application, before step S230, the claim settlement review method provided by the embodiments of the present application further comprises: adding a text mark at a text start part and a text end part of the sample disease information respectively to update the sample disease information, the text mark comprising a text start mark corresponding to the text start part and a text end mark corresponding to the text end part, the text start mark being used to represent that the sample disease information starts to be processed, and the text end mark being used to represent that the sample disease information ends to be processed.

[0102] Specifically, in order to quickly locate the sample disease information that needs to be generated by data augmentation in the model, a text start mark [cls] is added before the text start part of the sample disease information, the text start mark [cls] being used to represent that the sample disease information starts to be processed. A text end mark [sep] is added after the text end part of the sample disease information, the text end mark [sep] being used to represent that the sample disease information ends to be processed, so as to obtain the updated sample disease information. It should be noted that the sample disease information in the embodiments of the present application corresponds to the claim settlement disease information.

[0103] Please refer to Figure 3 , Figure 3 is a flowchart of the specific method of step S230 provided by the embodiments of the present application. In some embodiments, step S230 specifically comprises but is not limited to steps S310 to S350, which will be described below in combination with Figure 3 The five steps will be described in detail.

[0104] Step S310, input the sample disease information into the initial previous illness matching model, when the data augmentation sub-model detects the text start mark of the sample disease information, the sample disease information is processed for prediction to obtain first predicted text information;

[0105] Step S320, add the first predicted text information behind the text end mark of the sample disease information to obtain candidate sample information;

[0106] Step S330, when it is detected that the first predicted text information is not the text end mark, input the sample disease information and the first predicted text information into the initial previous illness matching model, process the sample disease information for prediction to obtain second predicted text information;

[0107] Step S340, add the second predicted text information behind the candidate sample information to update the candidate sample information;

[0108] Step S350, set the second predicted text information as the first predicted text information, and repeat the prediction processing according to the sample disease information and the first predicted text information until the first predicted text information is detected as a text end mark, add a text end mark behind the candidate sample information, and obtain target sample pair information.

[0109] In steps S310 to S350 of some embodiments, in order to improve the generalization ability of the previous disease matching model for processing of unobserved samples, the BERT algorithm is used to constitute a network structure of a data enhancement sub-model generated by data enhancement. The sample disease information is input into the initial previous disease matching model, and when the data enhancement sub-model detects a text start mark of the sample disease information, the first word information of the sample disease information is processed. Specifically, when a sentence segment1 input into the data enhancement sub-model needs to generate a sentence segment2, the initial Mask matrix corresponding to segment1 and segment2 is obtained by using the Mask matrix of the attention mechanism in the BERT model, wherein the mask indicates that the word information at this position is invisible to the model, and the non-mask indicates that the word information at this position is visible to the model, and by controlling the change of the mask in the initial Mask matrix, segment2 is generated word by word according to segment1.

[0110] In a specific embodiment, as shown in Table 1, the sample disease information "[cls] right palm cyst [sep]" input to the initial comorbidity matching model is input to the initial comorbidity matching model. In the first step, when the data enhancement sub-model detects the text start mark [cls] of the sample disease information, the sample disease information is processed for prediction, that is, the sample disease information is predicted word by word according to the Mask matrix of the attention mechanism of the model, and the first predicted text information "right" is obtained. Then, the first predicted text information "right" is added behind the text end mark of the sample disease information, and the candidate sample information "[cls] right palm cyst [sep] right" is obtained. In the second step, when it is detected that the first predicted text information is not the text end mark [sep], the sample disease information and the first predicted text information are input to the initial comorbidity matching model at the same time, and the sample disease information is processed for prediction, and the second predicted text information "right palm" is obtained. The second predicted text information "right palm" is added behind the candidate sample information, and the candidate sample information is updated, that is, the updated candidate sample information "[cls] right palm cyst [sep] right palm" is obtained. In the third step to the Nth step, the second predicted text information is set as the first predicted text information, and the prediction processing according to the sample disease information and the first predicted text information is repeated until it is detected that the first predicted text information is the text end mark, that is, the prediction of the sample disease information is completed, and finally the text end mark [sep] is added behind the candidate sample information, and the combination of the sample disease information and the target sample comorbidity information "[cls] right palm cyst [sep] right palm epidermal cyst for one year [sep]" is obtained. Then, the target sample pair information is obtained by combining the sample disease information with the corresponding sample label information.

[0111] Original input [cls] right palm cyst [sep] First step generation [cls] right palm cyst [sep] right Second step generation [cls] right palm cyst [sep] right palm ... ... N-1th step generation [cls] right palm cyst [sep] right palm epidermal cyst one year Nth step generation [cls] right palm cyst [sep] right palm epidermal cyst one year [sep]

[0112] Table 1

[0113] Please refer to Figure 4 , Figure 4 is a flowchart of the specific method of step S270 provided by the embodiments of the present application. In some embodiments, the model parameters of the initial comorbidity matching model are used to train the data enhancement sub-model and the comorbidity matching sub-model, and step S270 specifically includes but is not limited to steps S410 to S430, which will be described below in combination with Figure 4 The three steps will be described in detail.

[0114] Step S410, adjusting the model parameters of the comorbidity matching sub-model according to the label loss value;

[0115] Step S420, training the data enhancement sub-model according to the adjusted model parameters to update the second sample set;

[0116] In step S430, the adjusted initial previous illness matching model is continuously trained based on the second sample set until the label loss value meets the preset training end condition, to obtain the previous illness matching model.

[0117] In steps S410 to S430 of some embodiments, in order to improve the efficiency of the previous illness matching model in claim settlement review of the previous illness relationship, the data enhancement training of the model and the previous illness matching training adopt the same model structure, that is, the model parameters of the initial previous illness matching model are also used to train the data enhancement sub-model and the previous illness matching sub-model, so that the model parameters are shared. Specifically, the label loss value obtained according to the sample label information and the target label information is used to adjust the model parameters of the previous illness matching sub-model, and the data enhancement sub-model is trained according to the adjusted model parameters to update the second sample set. The adjusted initial previous illness matching model is continuously trained based on the second sample set until the label loss value meets the preset training end condition, to obtain the previous illness matching model.

[0118] The claim settlement review method provided by the embodiments of the present application is used to audit and identify the previous illness relationship of the current claim case. The claim settlement information of the user to be settled is obtained, and the claim settlement information includes claim disease information and to-be-audited previous illness information. In order to solve the problem that the accuracy of claim settlement review of the previous illness relationship is low due to different writing methods of the to-be-audited claim settlement information, the pre-trained previous illness matching model is obtained, the claim disease information is input into the previous illness matching model for data enhancement, the target disease information is obtained, the target disease information includes the claim disease information and target previous illness information, and the target previous illness information is used to update the training of the previous illness matching model. Then, the previous illness matching model is used to perform previous illness matching on the to-be-audited claim settlement information, to obtain the causal label information, and the causal label information is used to represent the causal relationship between the claim disease information and the to-be-audited previous illness information. Finally, the claim settlement result of the user to be settled is determined according to the causal label information. When the causal label information indicates that the claim disease information and the to-be-audited previous illness information have a causal relationship, it is determined that the claim settlement result is to cancel the claim settlement; and when the causal label information indicates that the claim disease information and the to-be-audited previous illness information do not have a causal relationship, it is determined that the claim settlement result is to agree to the claim settlement. The present application introduces data enhancement in the model, enriches the data samples in the training stage, and adopts the same model structure for the data enhancement training of the model and the previous illness matching training, that is, the model parameters of the initial previous illness matching model are also used to train the data enhancement sub-model and the previous illness matching sub-model, so that the model parameters are shared, and the efficiency and accuracy of the claim settlement review of the previous illness relationship are improved.

[0119] Please refer to Figure 5 , Figure 5is a module structure block diagram of the claim auditing apparatus provided in the embodiment of the present application. In some embodiments, the claim auditing apparatus comprises an information acquisition module 510, a model acquisition module 520, a data enhancement module 530, a previous illness matching module 540, and a claim determination module 550.

[0120] The information acquisition module 510 is configured to acquire claim information to be audited of a user to be claimed, wherein the claim information to be audited comprises claim disease information and previous illness information to be audited.

[0121] The model acquisition module 520 is configured to acquire a pre-trained previous illness matching model.

[0122] The data enhancement module 530 is configured to input the claim disease information into the previous illness matching model for data enhancement, to obtain target disease information, wherein the target disease information comprises the claim disease information and target previous illness information, and the target previous illness information is used to update the training previous illness matching model.

[0123] The previous illness matching module 540 is configured to perform previous illness matching on the claim information to be audited according to the previous illness matching model, to obtain cause-effect label information, wherein the cause-effect label information is used to represent a cause-effect relationship between the claim disease information and the previous illness information to be audited.

[0124] The claim determination module 550 is configured to determine a claim result of the user to be claimed according to the cause-effect label information.

[0125] In some embodiments, the claim determination module 550 of the claim auditing apparatus provided in the embodiment of the present application specifically comprises a first claim result submodule or a second claim result submodule.

[0126] The first claim result submodule is configured to, when the cause-effect label information indicates that the claim disease information and the previous illness information to be audited have a cause-effect relationship, determine that the claim result is to cancel the claim.

[0127] The second claim result submodule is configured to, when the cause-effect label information indicates that the claim disease information and the previous illness information to be audited do not have a cause-effect relationship, determine that the claim result is to approve the claim.

[0128] It should be noted that the claim auditing apparatus in the embodiment of the present application is used to execute the claim auditing method in the above-mentioned embodiments, and the specific processing process is the same as that of the claim auditing method in the above-mentioned embodiments, which will not be repeated here.

[0129] The embodiment of the present application also provides a computer device comprising a memory and a processor, wherein the memory stores a computer program, and the computer program is executed by the processor to execute the claim auditing method in the above-mentioned embodiments of the present application.

[0130] The computer device provided in the embodiment of the present application is used for executing the claim settlement auditing method, and the specific processing procedure is the same as that of the claim settlement auditing method in the above embodiment, which will not be repeated here.

[0131] The following will be described in combination with Figure 6 The hardware structure of the computer device will be described in detail. The computer device comprises a processor 610, a memory 620, an input / output interface 630, a communication interface 640 and a bus 650.

[0132] The processor 610 can be implemented in the form of a general CPU (Central Processing Unit, central processor), a microprocessor, an ASIC (Application Specific Integrated Circuit, application specific integrated circuit) or one or more integrated circuits, and is used for executing related programs to implement the technical solutions provided by the embodiments of the present application.

[0133] The memory 620 can be implemented in the form of a ROM (Read Only Memory, read-only memory), a static storage device, a dynamic storage device or a RAM (Random Access Memory, random access memory). The memory 620 can store an operating system and other application programs. When the technical solutions provided by the embodiments of the present application are implemented by software or firmware, the related program codes are stored in the memory 620 and called and executed by the processor 610 to implement the claim settlement auditing method of the embodiments of the present application.

[0134] The input / output interface 630 is used for realizing information input and output.

[0135] The communication interface 640 is used for realizing the communication interaction between the device and other devices. The communication can be realized by a wired manner (for example, USB, network cable, etc.) or a wireless manner (for example, mobile network, WIFI, Bluetooth, etc.).

[0136] The bus 650 is used for transmitting information between various components (for example, the processor 610, the memory 620, the input / output interface 630 and the communication interface 640) of the device.

[0137] The processor 610, the memory 620, the input / output interface 630 and the communication interface 640 are connected with each other through the bus 650 to realize the communication connection between them in the device.

[0138] The embodiment of the present application further provides a storage medium, which is a computer readable storage medium and stores a computer program. When the computer program is executed by a computer, the computer is used to execute the claim settlement auditing method of the embodiments of the present application.

[0139] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. In addition, the memory can include a high-speed random access memory and can also include a non-transitory memory, such as at least one magnetic disk storage device, a flash memory device, or other non-transitory solid-state memory device. In some embodiments, the memory can optionally include a memory that is remotely located with respect to the processor, and these remote memories can be connected to the processor through a network. Examples of the above network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0140] The embodiments described in the embodiments of the present application are for more clearly illustrating the technical solutions of the embodiments of the present application, and do not constitute a limitation on the technical solutions provided by the embodiments of the present application. Those skilled in the art can know that, with the evolution of technology and the appearance of new application scenarios, the technical solutions provided by the embodiments of the present application are also applicable to similar technical problems.

[0141] Those skilled in the art can understand that, Figures 1 to 4 The technical solutions shown in the above description do not constitute a limitation on the embodiments of the present application, and can include more or fewer steps than shown in the figures, or combine certain steps, or different steps.

[0142] The device embodiments described above are only schematic, and units described as separate components can or can not be physically separate, i.e., can be located in one place or distributed over multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the embodiments of the present application.

[0143] Those skilled in the art can understand that all or some of the steps in the above disclosed method, the functional modules / units in the system and the device can be implemented as software, firmware, hardware and their appropriate combinations.

[0144] The terms "first", "second", "third", "fourth" and the like used in the description of the present application and the above drawings, if any, are used to distinguish similar objects, and do not necessarily indicate a specific order or sequence. It should be understood that the data thus used can be interchanged under appropriate circumstances, so that the embodiments of the present application described herein can be implemented in other than the order illustrated or described herein. In addition, the terms "include" and "have" and any variations thereof are intended to cover non-exclusive inclusion, for example, a process, method, system, product or device that includes a series of steps or units does not necessarily have to be limited to those steps or units clearly listed, but can include other steps or units not clearly listed or inherent to these processes, methods, products or devices.

[0145] It should be understood that, in the application, "at least one" means one or more, "multiple" means two or more. "And / or" is used to describe the relationship between the associated objects, which means that there can be three relationships, for example, "A and / or B" can represent three cases: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally represents an "or" relationship between the associated objects. "At least one of the following" or similar expressions means any combination of these items, including single or multiple combinations of any combination. For example, at least one of a, b or c can represent: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, and c can be single or multiple.

[0146] In several embodiments provided in the application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only illustrative, for example, the division of the units is only a logical function division, and actual implementation can have another division manner, for example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. In addition, the coupling or direct coupling or communication connection between the displayed or discussed units can be indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0147] The units described as separate components can or can not be physically separated, and the components displayed as units can or can not be physical units, that is, they can be located in one place, or they can be distributed on multiple network units. According to actual needs, part or all of the units can be selected to achieve the purpose of the embodiment scheme.

[0148] In addition, each functional unit in each embodiment of the application can be integrated into a processing unit, or each unit can exist physically, or two or more units can be integrated into one unit. The integrated unit can be realized in the form of hardware or in the form of a software functional unit.

[0149] The integrated unit, if implemented in the form of a software function unit and sold or used as an independent product, can be stored in a computer readable storage medium. Based on such understanding, the technical solutions of the present application, essentially or in other words, the part that contributes to the prior art or the whole or part of the technical solutions can be embodied in the form of a software product. The computer software product is stored in a storage medium, and includes multiple instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The aforementioned storage medium includes: a U disk, a mobile hard disk, a read-only memory (ROM), a random access memory (RAM), a magnetic disk or an optical disk, and various program storage media.

[0150] The preferred embodiments of the embodiments of the present application are described above with reference to the accompanying drawings, and are not limited to the scope of the embodiments of the present application. Any modifications, equivalent replacements and improvements made by those skilled in the art without departing from the scope and essence of the embodiments of the present application shall be within the scope of the embodiments of the present application.

Claims

1. A method of claim settlement auditing, the method comprising: The method comprises: obtaining the to-be-audited claim information of the to-be-claimed user, wherein the to-be-audited claim information comprises claim disease information and to-be-audited past medical history information; obtaining a pre-trained past medical history matching model; the pre-trained past medical history matching model is obtained by training in the following manner: constructing a first sample set, wherein the first sample set comprises a plurality of first sample pair information, the first sample pair information comprises sample disease information, first sample past medical history information and sample label information, and the sample label information is used to represent the causal relationship between the sample disease information and the first sample past medical history information; constructing an initial past medical history matching model by using a BERT algorithm, wherein the initial past medical history matching model comprises a data enhancement sub-model and a past medical history matching sub-model; adding a text mark at the beginning and end of the text of the sample disease information to update the sample disease information, wherein the text mark comprises a text start mark corresponding to the beginning of the text and a text end mark corresponding to the end of the text, the text start mark is used to represent the start of processing of the sample disease information, and the text end mark is used to represent the end of processing of the sample disease information; inputting the sample disease information into the initial past medical history matching model, performing data enhancement on the sample disease information according to the data enhancement sub-model to obtain target sample pair information, wherein the target sample pair information is the combination of the sample disease information and target sample past medical history information predicted based on the text start mark and the text end mark; constructing a second sample set according to the first sample set and the target sample pair information, and training the initial past medical history matching model according to the second sample set to obtain the past medical history matching model; inputting the claim disease information into the past medical history matching model for data enhancement to obtain target disease information, wherein the target disease information comprises the claim disease information and target past medical history information, and the target past medical history information is used to update the training of the past medical history matching model; performing past medical history matching on the to-be-audited claim information according to the past medical history matching model to obtain causal label information, wherein the causal label information is used to represent the causal relationship between the claim disease information and the to-be-audited past medical history information; determining the claim result of the to-be-claimed user according to the causal label information.

2. The method of claim 1, wherein, The determination of the claim result of the to-be-claimed user according to the causal label information comprises: when the causal label information indicates that the claim disease information and the to-be-audited past medical history information have a causal relationship, determining that the claim result is to cancel the claim; when the causal label information indicates that the claim disease information and the to-be-audited past medical history information do not have a causal relationship, determining that the claim result is to approve the claim.

3. The method according to claim 1 or 2, characterized in that, The target sample pair information includes the sample disease information, target sample past medical history information, and the sample label information, the second sample set includes a plurality of second sample pair information, the second sample pair information is the first sample pair information or the target sample pair information, the second sample pair information includes the sample disease information, second past medical history information, and the sample label information, the second past medical history information is the first sample past medical history information or the target sample past medical history information, and the initial past medical history matching model is trained according to the second sample set to obtain the past medical history matching model, including: performing past medical history matching on the sample disease information and the second past medical history information according to the past medical history matching sub-model to obtain target label information; determining a label loss value according to the sample label information and the target label information; adjusting model parameters of the initial past medical history matching model according to the label loss value, and continuing to train the adjusted initial past medical history matching model based on the second sample set until the label loss value meets a preset training end condition, so as to obtain the past medical history matching model.

4. The method of claim 1, wherein, The sample disease information is input into the initial past medical history matching model, and the data augmentation sub-model is used to perform data augmentation on the sample disease information to obtain target sample pair information, including: the sample disease information is input into the initial past medical history matching model, and when the data augmentation sub-model detects the text start mark of the sample disease information, the sample disease information is predicted and processed to obtain first predicted text information; the first predicted text information is added behind the text end mark of the sample disease information to obtain candidate sample information; when it is detected that the first predicted text information is not the text end mark, the sample disease information and the first predicted text information are input into the initial past medical history matching model, the sample disease information is predicted and processed to obtain second predicted text information; the second predicted text information is added behind the candidate sample information to update the candidate sample information; the second predicted text information is set as the first predicted text information, and the prediction and processing according to the sample disease information and the first predicted text information is repeated until, when it is detected that the first predicted text information is the text end mark, the text end mark is added behind the candidate sample information to obtain target sample pair information.

5. The method of claim 3, wherein, The model parameters of the initial past medical history matching model are used to train the data augmentation sub-model and the past medical history matching sub-model, the model parameters of the initial past medical history matching model are adjusted according to the label loss value, and the adjusted initial past medical history matching model is continuously trained based on the second sample set until the label loss value meets a preset training end condition, so as to obtain the past medical history matching model, including: adjusting model parameters of the past medical history matching sub-model according to the label loss value; training the data augmentation sub-model according to the adjusted model parameters to update the second sample set; continuing to train the adjusted initial past illness matching model based on the second sample set until the label loss value meets a preset training end condition, to obtain the past illness matching model.

6. A claim settlement auditing apparatus characterized by comprising: The device comprises: An information acquisition module is configured to acquire to-be-audited claim information of a to-be-claimed user, wherein the to-be-audited claim information comprises claim disease information and to-be-audited past illness information. A model acquisition module is configured to acquire a pre-trained past illness matching model, wherein the pre-trained past illness matching model is obtained by training in the following manner: constructing a first sample set, wherein the first sample set comprises a plurality of first sample pair information, the first sample pair information comprises sample disease information, first sample past illness information, and sample label information, and the sample label information is used to represent a causal relationship between the sample disease information and the first sample past illness information; constructing an initial past illness matching model by using a BERT algorithm, wherein the initial past illness matching model comprises a data augmentation sub-model and a past illness matching sub-model; adding a text mark at a text start part and a text end part of the sample disease information respectively to update the sample disease information, wherein the text mark comprises a text start mark corresponding to the text start part and a text end mark corresponding to the text end part, the text start mark is used to represent that the sample disease information is processed, and the text end mark is used to represent that the sample disease information is processed; inputting the sample disease information into the initial past illness matching model, performing data augmentation on the sample disease information according to the data augmentation sub-model to obtain target sample pair information, wherein the target sample pair information is combined information of the sample disease information and target sample past illness information predicted based on the text start mark and the text end mark; constructing a second sample set according to the first sample set and the target sample pair information, and training the initial past illness matching model according to the second sample set to obtain the past illness matching model; A data augmentation module is configured to input the claim disease information into the past illness matching model to perform data augmentation, to obtain target disease information, wherein the target disease information comprises the claim disease information and target past illness information, and the target past illness information is used to update the past illness matching model. A past illness matching module is configured to perform past illness matching on the to-be-audited claim information according to the past illness matching model, to obtain causal label information, wherein the causal label information is used to represent a causal relationship between the claim disease information and the to-be-audited past illness information. A claim determination module is configured to determine a claim result of the to-be-claimed user according to the causal label information.

7. The apparatus of claim 6, wherein, The claim determination module is configured to determine a claim result of the to-be-claimed user according to the causal label information, comprising: A first claim result submodule is configured to determine that the claim result is claim cancellation when the causal label information indicates that the claim disease information and the to-be-audited past illness information have a causal relationship. A second claim result sub-module is configured to determine the claim result as approval of the claim when the cause-effect label information indicates that the claim disease information and the to-be-audited previous illness information do not have a cause-effect relationship.

8. A computer device, comprising: The computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor is configured to execute the computer program to perform the following steps: The method of any one of claims 1 to 5.

9. A storage medium, the storage medium being a computer-readable storage medium, characterized in that, The computer readable storage medium stores a computer program, and the computer is configured to execute the computer program to perform the following steps: The method of any one of claims 1 to 5. The computer readable storage medium stores a computer program, and the computer is configured to execute the computer program to perform the following steps:

Images