Claim settlement resource prediction method and device, electronic device, and storage medium

Abstract

This application provides a method, apparatus, electronic device, and storage medium for predicting claims resources, belonging to the field of artificial intelligence technology and applicable to the financial technology field. The method includes: obtaining the initial policy text of a target object; obtaining target claims data based on a target claims request; filtering the initial policy text based on the target claims data to obtain target policy text; parsing the target policy text using a large-scale text parsing model to obtain policy attribute data; obtaining a claims resource prediction function based on the target claims request; selecting target claims attribute data matching preset claims attribute parameters from the policy attribute data; adjusting the claims resource prediction function based on the target claims attribute data to obtain a target resource prediction function for the target object; and predicting claims resources for the target claims data based on the target resource prediction function to obtain the target claims resources. This application can improve the accuracy and efficiency of claims resource prediction.

Description

Technical Field

[0001] This application relates to the field of artificial intelligence technology, and is applicable to the financial technology field, particularly to a method and apparatus for predicting claims resources, an electronic device, and a storage medium. Background Technology

[0002] Claims resources refer to the amount of resources that a resource provider (such as an insurance company or traffic provider) pays out based on a claimant's claim. For example, in the insurance sector of fintech, claims resources can be the premiums paid out to a claimant. Because each claimant's choice of claim items and their own circumstances differ, the specific claims information used in calculating claims resources, such as deductibles, payout ratios, and payout caps, varies.

[0003] Currently, when determining the corresponding claim resources for a claimant, the relevant technology typically involves claims staff manually identifying the necessary claim information for each claimant during resource calculation and inputting this information into the claims system for resource prediction. However, this method requires claims staff to possess in-depth insurance knowledge and a precise understanding of policy terms. Furthermore, due to the low efficiency and susceptibility to errors inherent in manual operation, it can easily affect the efficiency and accuracy of claim resource prediction. Therefore, improving the accuracy and efficiency of claim resource prediction has become a pressing technical problem to be solved. Summary of the Invention

[0004] The main objective of this application is to provide a method, apparatus, electronic device, and storage medium for predicting claims resources, aiming to improve the accuracy and efficiency of claims resource prediction.

[0005] To achieve the above objectives, a first aspect of this application proposes a method for predicting claims resources, the method comprising:

[0006] Retrieve the initial policy text of the target object;

[0007] In response to the target claim request of the target object, and based on the target claim request, obtain the target claim data of the target object;

[0008] Based on the target claims data, the initial policy text is filtered to obtain the target policy text;

[0009] Based on a pre-defined text parsing model, the target policy text is parsed to obtain policy attribute data;

[0010] A claim resource prediction function is obtained based on the target claim request, and the claim resource prediction function includes preset claim attribute parameters;

[0011] Select target claim attribute data that matches the preset claim attribute parameters from the policy attribute data;

[0012] Based on the target claims attribute data, the claims resource prediction function is adjusted to obtain the target resource prediction function for the target object;

[0013] Based on the target resource prediction function, the target claims data is used to predict claims resources to obtain target claims resources.

[0014] In some embodiments, the step of predicting the target claims resources based on the target claims data using the target resource prediction function to obtain the target claims resources includes:

[0015] Based on the target policy text, the function prediction conditions of the target resource prediction function are determined. The function prediction conditions include prediction condition parameters. The function prediction conditions are used to indicate the conditions that the data in the target claims data that matches the prediction condition parameters must meet when the target resource prediction function is called.

[0016] Based on the predicted condition parameters, the target claim data is matched to obtain claim condition data;

[0017] Based on the function prediction conditions, the claim condition data is inspected to obtain the function state of the target resource prediction function. The function state is used to indicate that the target condition data that matches the prediction condition parameters satisfies the function prediction conditions.

[0018] Based on the target resource prediction function and the function state, the target claim data is used to predict the claim resources, thereby obtaining the target claim resources.

[0019] In some embodiments, the function state includes a call state, which is used to indicate that the target condition data matching the prediction condition parameters satisfies the function prediction conditions, and the claims resource prediction function further includes preset claims object parameters;

[0020] The step of predicting claim resources based on the target claim data using the target resource prediction function and the function state to obtain the target claim resources includes:

[0021] If the function is in the calling state, the target claim data is matched based on the preset claim object parameters to obtain the claim object data;

[0022] Based on the target resource prediction function and the claims object data, the claims resource is predicted to obtain the target claims resource.

[0023] In some embodiments, adjusting the claims resource prediction function based on the target claims attribute data to obtain the target resource prediction function for the target object includes:

[0024] Based on the target claim attribute data, attribute parameter division intervals are obtained from preset structured claim data, and the attribute parameter division intervals match the preset claim attribute parameters;

[0025] Based on the attribute parameters, the function range of the claims resource prediction function is adjusted to obtain the candidate resource prediction function.

[0026] The target resource prediction function is obtained by adjusting the function parameters of the candidate resource prediction function based on the target claim attribute data.

[0027] In some embodiments, after obtaining the attribute parameter division range from the preset structured claims data based on the target claims attribute data, the method further includes:

[0028] The preset claims attribute parameters are encoded to obtain claims parameter codes;

[0029] The target claim attribute data and the preset claim clause categories are matched to obtain the target claim clause category;

[0030] The attribute parameters that match the target claim attribute data are divided into intervals, and the claim parameter codes are combined with interval codes to obtain structured attribute data;

[0031] The structured attribute data that matches the target claim terms category are combined to obtain the structured claim data for the target object.

[0032] In some embodiments, obtaining the initial policy text of the target object includes:

[0033] Obtain the object account information of the target object;

[0034] Based on the object's account information, the original policy data is matched from a preset database;

[0035] The original policy data is subjected to text recognition to obtain the original policy text;

[0036] The original policy text is preprocessed to obtain the initial policy text.

[0037] In some embodiments, the method further includes:

[0038] Obtain the object claim level of the target object;

[0039] Obtain the resource review model that matches the claim level of the object;

[0040] Based on the resource review model, the target claim resources are reviewed to obtain the resource status of the target claim resources. The resource status is used to indicate whether the target claim resources have passed the review or failed the review.

[0041] If the resource status indicates that the target claim resource has been approved, the target claim resource will be transferred.

[0042] To achieve the above objectives, a second aspect of this application provides a claims resource prediction apparatus, the apparatus comprising:

[0043] The first acquisition module is used to acquire the initial policy text of the target object;

[0044] The second acquisition module is used to respond to the target claim request of the target object and acquire the target claim data of the object based on the target claim request;

[0045] The text filtering module is used to filter the initial policy text based on the target claims data to obtain the target policy text.

[0046] The parsing module is used to parse the target policy text based on a preset text parsing model to obtain policy attribute data;

[0047] The third acquisition module is used to acquire a claim resource prediction function based on the target claim request, wherein the claim resource prediction function includes preset claim attribute parameters;

[0048] The data matching module is used to select target claim attribute data that matches the preset claim attribute parameters from the policy attribute data;

[0049] The function adjustment module is used to adjust the claim resource prediction function based on the target claim attribute data to obtain the target resource prediction function for the target object.

[0050] The resource prediction module is used to predict the claim resources of the target claim data based on the target resource prediction function, so as to obtain the target claim resources.

[0051] To achieve the above objectives, a third aspect of this application provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the claims resource prediction method described in the first aspect.

[0052] To achieve the above objectives, a fourth aspect of the present application provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the claims resource prediction method described in the first aspect.

[0053] This application proposes a method, apparatus, electronic device, and storage medium for predicting claims resources. The method involves: acquiring the initial policy text of a target object; responding to a target claims request from the target object and acquiring target claims data based on the request; further, filtering the initial policy text based on the target claims data to obtain the target policy text; parsing the target policy text based on a pre-defined text parsing model to obtain policy attribute data; further, acquiring a claims resource prediction function based on the target claims request, the function including pre-defined claims attribute parameters; selecting target claims attribute data matching the pre-defined parameters from the policy attribute data; further, adjusting the claims resource prediction function based on the target claims attribute data to obtain the target resource prediction function for the target object; and finally, predicting claims resources based on the target claims data using the target resource prediction function to obtain the target claims resources. Compared to related technologies that use uniform attribute parameters for claims resource calculation, or that manually identify the claims information required for each claimant to update the claims calculation, this application can obtain policy attribute data from the target policy text of the target object based on the target claims data. It then adjusts the claims resource prediction function based on the target claims attribute data that matches the policy attribute information with preset claims attribute parameters, thereby accurately predicting the target claims resources of the target object. Thus, this application employs an automated attribute data extraction method, which can accurately adjust the function to obtain the resource prediction function for each object, effectively improving the accuracy and efficiency of claims resource prediction. Attached Figure Description

[0054] Figure 1 This is the first flowchart of the claims resource prediction method provided in the embodiments of this application;

[0055] Figure 2 yes Figure 1 A flowchart of step S110 in the process;

[0056] Figure 3 yes Figure 1 A flowchart of step S170 in the process;

[0057] Figure 4 This is the second flowchart of the claims resource prediction method provided in the embodiments of this application;

[0058] Figure 5 yes Figure 1 A flowchart of step S180 in the process;

[0059] Figure 6 yes Figure 5 A flowchart of step S540 in the process;

[0060] Figure 7 This is the third flowchart of the claims resource prediction method provided in the embodiments of this application;

[0061] Figure 8 This is a schematic diagram of a claims resource prediction device provided in an embodiment of this application;

[0062] Figure 9 This is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of this application. Detailed Implementation

[0063] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0064] It should be noted that although functional modules are divided in the device schematic diagram and a logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first," "second," etc., in the specification, claims, and the aforementioned drawings are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence.

[0065] 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 herein is for the purpose of describing embodiments of this application only and is not intended to limit this application.

[0066] First, let's analyze some of the terms used in this application:

[0067] Artificial intelligence (AI) is a new branch of computer science that studies, develops, and applies theories, methods, technologies, and systems to simulate, extend, and expand human intelligence. It aims to understand the essence of intelligence and produce intelligent machines that can react in a way similar to human intelligence. Research in this field includes robotics, speech recognition, image recognition, natural language processing, and expert systems. AI can simulate the information processes of human consciousness and thought. Furthermore, AI utilizes digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceiving the environment, acquiring knowledge, and using that knowledge to achieve optimal results.

[0068] Claims resources refer to the amount of resources that a resource provider (such as an insurance company or traffic provider) pays out based on a claimant's claim. For example, in the insurance sector of fintech, claims resources can be the premiums paid out to a claimant. Because each claimant's choice of claim items and their own circumstances differ, the specific claims information used in calculating claims resources, such as deductibles, payout ratios, and payout caps, varies.

[0069] Currently, when determining the corresponding claim resources for a claimant, the relevant technology typically involves claims staff manually identifying the necessary claim information for each claimant's resource calculation and inputting this information into the claims system for resource prediction. However, this method requires claims staff to possess in-depth insurance knowledge and a precise understanding of policy terms, enabling them to carefully read policy terms and special conditions to determine the applicable calculation rules. Because manual operation is inefficient and prone to errors, it can easily affect the efficiency and accuracy of claim resource prediction. Furthermore, when dealing with a large number of policies and complex claim conditions, this manual method severely limits claims efficiency and service quality. Therefore, improving the accuracy and efficiency of claim resource prediction has become an urgent technical problem to be solved.

[0070] Based on this, embodiments of this application provide a method and apparatus, electronic device, and storage medium for predicting claims resources, which can improve the accuracy and efficiency of predicting claims resources.

[0071] The embodiments of this application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence (AI) refers to the theories, methods, technologies, and application systems that use digital computers or machines controlled by digital computers to simulate, extend, and expand human intelligence, perceive the environment, acquire knowledge, and use that knowledge to obtain optimal results.

[0072] Foundational technologies for artificial intelligence generally include sensors, dedicated AI chips, cloud computing, distributed storage, big data processing, operating / interactive systems, and mechatronics. AI software technologies mainly encompass computer vision, robotics, biometrics, speech processing, natural language processing, and machine learning / deep learning.

[0073] The claims resource prediction method provided in this application relates to the field of artificial intelligence technology. This method can be applied to a terminal, a server, or software running on either a terminal or a server. In some embodiments, the terminal can be a smartphone, tablet, laptop, desktop computer, etc.; the server can be configured as an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server providing basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, content delivery networks (CDN), and big data and artificial intelligence platforms; the software can be an application implementing the claims resource prediction method, but is not limited to the above forms.

[0074] This application can be used in a wide variety of general-purpose or special-purpose computer system environments or configurations. Examples include: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, and distributed computing environments including any of the above systems or devices. This application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc., that perform specific tasks or implement specific abstract data types. This application can also be practiced in distributed computing environments where tasks are performed by remote processing devices connected via a communication network. In distributed computing environments, program modules can reside in local and remote computer storage media, including storage devices.

[0075] It should be noted that in all specific embodiments of this application, when processing data related to the identity or characteristics of an object, such as the object's policy information, claims data, and attribute characteristics, the object's permission or consent is obtained first. Furthermore, the collection, use, and processing of this data comply with relevant laws, regulations, and standards. In addition, when embodiments of this application require access to sensitive personal information of an object, separate permission or consent from the object is obtained through pop-ups or redirection to a confirmation page. Only after obtaining the object's separate permission or consent is the necessary object-related data required for the proper functioning of the embodiments of this application obtained.

[0076] Please see Figure 1 , Figure 1 This is an optional flowchart of the claims resource prediction method provided in the embodiments of this application. In some embodiments, Figure 1 The method may include, but is not limited to, steps S110 to S180:

[0077] Step S110: Obtain the initial policy text of the target object;

[0078] Step S120: In response to the target claim request of the target object, obtain the target claim data of the target object based on the target claim request;

[0079] Step S130: Based on the target claims data, perform text filtering on the initial policy text to obtain the target policy text;

[0080] Step S140: Based on the preset text parsing model, perform text parsing on the target policy text to obtain policy attribute data;

[0081] Step S150: Obtain the claim resource prediction function based on the target claim request;

[0082] Step S160: Select the target claim attribute data that matches the preset claim attribute parameters from the policy attribute data;

[0083] Step S170: Adjust the claim resource prediction function based on the target claim attribute data to obtain the target resource prediction function for the target object;

[0084] Step S180: Based on the target resource prediction function, predict the target claims data to obtain the target claims resources.

[0085] In step S110 of some embodiments, the initial policy text may refer to the textual form of the policy data that the target object participates in and signs. For example, in the field of insurance in fintech, the target object is the person who purchases a specific insurance, and the initial policy text corresponds to the insurance contract document signed by the target object. The initial policy text may be a PDF, Word document, or other format file, containing all the terms and conditions of claims related to the insurance participated in by the target object.

[0086] It should be noted that due to differences in each target individual's choice of claim items and their own circumstances, the specific claim information used in calculating claim resources, such as deductibles, reimbursement ratios, and reimbursement caps, will vary. For example, if a target individual is a first-time insured and does not have a long history of hospitalizations, the reimbursement ratio and reimbursement cap used in calculating claim resources will be set relatively high. If a target individual is an nth-time insured (n is a positive integer greater than 1) and has a lot of long hospitalizations, the reimbursement ratio and reimbursement cap used in calculating claim resources may decrease as the number of insured events increases. Therefore, using a fixed claim resource calculation method or manually identifying relevant information for calculation is not suitable for the current practical scenario.

[0087] See Figure 2 , Figure 2 This is an optional flowchart of step S110 provided in the embodiments of this application. In some embodiments of this application, step S110 may specifically include steps S210 to S240:

[0088] Step S210: Obtain the object account information of the target object;

[0089] Step S220: Match the original policy data from the preset database based on the object account information;

[0090] Step S230: Perform text recognition on the original policy data to obtain the original policy text;

[0091] Step S240: Perform text preprocessing on the original policy text to obtain the initial policy text.

[0092] In step S210 of some embodiments, this application can retrieve the initial policy text related to the target object from a preset database or storage system corresponding to the claims system. Specifically, the target object's account information can be obtained first. This account information may include at least one of the target object's insurance account, identity verification, or contact information.

[0093] In step S220 of some embodiments, this application can utilize the target object's account information to retrieve original policy data related to the target object from a preset database corresponding to the claims system. This original policy data can be an electronic version of the insurance contract signed by the target object, such as a scanned or photographed image of the insurance contract, a PDF document, etc., without limitation. The original policy data may include key information such as the specific terms of the insurance contract, the scope of insurance, and the sum insured, which is important for subsequent claims resource prediction.

[0094] In step S230 of some embodiments, in order to extract and process key information in the policy data in the future, this application can use optical character recognition (OCR) to realize text recognition, so as to convert the paper or image file of the policy into an editable text format, which can ensure that the policy content can be electronically processed and analyzed.

[0095] In step S240 of some embodiments, after obtaining the original policy text through text recognition, this application can perform text preprocessing on the original policy text. The text preprocessing process can include removing irrelevant information, formatting the text, correcting recognition errors, and filtering noise to ensure the accuracy and usability of the text. In this way, the obtained initial policy text can be used as the basis for subsequent claims resource prediction.

[0096] In steps S210 to S240 above, this application can query the original policy data and perform text recognition and text preprocessing on the original policy data to provide a text basis for subsequent claims resource prediction and improve the accuracy of claims information extraction.

[0097] In step S120 of some embodiments, when a target object needs to make a claim, it can submit a target claim request. At this time, the claims system can respond to the target claim request submitted by the target object and obtain the target claim data of the target object based on the target claim request. The target claim data refers to claim data related to the target claim request, such as accident details, loss details, hospitalization records, and records of relevant claim resources.

[0098] In step S130 of some embodiments, after obtaining the target claim data, the target claim data can be used to filter the portion of the initial policy text related to the current target claim request. For example, if the target claim request involves a specific type of loss, the system will filter out clauses related to that type of claim loss from the initial policy text. The target policy text is used to represent the policy text associated with the target claim request. That is, the initial policy text corresponding to the target object can correspond to multiple claim types, such as health insurance, car insurance, etc. If the target object needs to claim compensation for its own vehicle, then the target claim request is a claim request based on car insurance. In this case, it is necessary to filter out the policy text related to car insurance from the initial policy text, thus obtaining the target policy text.

[0099] In step S140 of some embodiments, in order to gain a deeper understanding of the relevant information on claims resource prediction in the target policy text, this application can utilize a preset text parsing model to parse the filtered target policy text and extract key policy attribute data, such as expected insurance resource amount, deductible, and insurance period. In other words, this application can use a pre-trained text parsing model to quickly scan the target policy text, identify and label all relevant claims attributes. The text parsing model can be based on historical data and real-world cases, and can employ different models, such as machine learning models (e.g., random forests, support vector machines), deep learning models, large-scale models, etc., without specific limitations.

[0100] In step S150 of some embodiments, since the claim resource prediction algorithms for specific loss types need to follow the same rules when predicting claim resources, this application can first obtain a claim resource prediction function based on the target claim request. This claim resource prediction function may include preset claim attribute parameters and preset claim object parameters. Specifically, the claim resource prediction function is a function used to predict the claim resources corresponding to the target claim request; the preset claim attribute parameters are attribute parameters related to the target policy text in the claim resource prediction function; and the preset claim object parameters are object parameters related to the target claim data in the claim resource prediction function. For example, if the target claim request is a disease claim, and the obtained claim resource prediction function is "Predicted claim resource = (Number of hospitalization days - Deductible days) × Payout ratio × Daily resource quantity," then the preset claim attribute parameters may include "Deductible days," "Payout ratio," and "Daily resource quantity," and the preset claim object parameter is "Number of hospitalization days."

[0101] In step S160 of some embodiments, in order to set personalized resource prediction functions for different target objects, this application can select target claim attribute data that matches preset claim attribute parameters from policy attribute data. This effectively reduces the error rate caused by human interpretation, improves the accuracy of claims calculation results, and achieves automated and personalized claims resource prediction.

[0102] In step S170 of some embodiments, after determining the target claim attribute data, this application can adjust the claim resource prediction function based on the target claim attribute data to adapt to the situation of a specific customer and a specific claim request, thereby obtaining a more accurate target resource prediction function.

[0103] See Figure 3 , Figure 3 This is an optional flowchart of step S170 provided in the embodiments of this application. In some embodiments of this application, step S170 may specifically include steps S310 to S330:

[0104] Step S310: Obtain the attribute parameter division range from the preset structured claims data based on the target claims attribute data;

[0105] Step S320: Adjust the function range of the claims resource prediction function based on the interval division of attribute parameters to obtain the candidate resource prediction function;

[0106] Step S330: Adjust the function parameters of the candidate resource prediction function based on the target claim attribute data to obtain the target resource prediction function for the target object.

[0107] In step S310 of some embodiments, since the target policy text does not simply disclose the numerical values ​​of the target claim attribute data, but may contain some special limiting requirements to define the range of attribute parameters, and different resource prediction methods can be used for data in different ranges, this application can first obtain the attribute parameter division range from the preset structured claim data based on the target claim attribute data, and the attribute parameter division range matches the preset claim attribute parameters. The structured claim data is a structured data format based on the relevant attribute parameters stored in the initial policy text of the target object, which facilitates quick querying of the attribute parameter division range corresponding to the target claim attribute data in the target policy text.

[0108] For example, the relevant settings for the preset claim attribute data of "deductible and compensation ratio" in the initial insurance policy text are as follows: "The deductible and compensation ratio are determined based on the severity of the insurance accident and the medical expenses. If the medical expenses are below X1, the deductible for accidental medical expenses is A1, and the compensation ratio for accidental medical expenses is B1; the deductible for disease hospitalization medical expenses is C1, and the compensation ratio for disease hospitalization medical expenses is D1. If the medical expenses are between X2 and X3 (X1 < X2 < X3), the deductible for accidental medical expenses is A2, and the compensation ratio for accidental medical expenses is B2; the deductible for disease hospitalization medical expenses is C2, and the compensation ratio for disease hospitalization medical expenses is D2. If the medical expenses are above X3, the deductible for accidental medical expenses is A3, and the compensation ratio for accidental medical expenses is B3; the deductible for disease hospitalization medical expenses is C3, and the compensation ratio for disease hospitalization medical expenses is D3." Thus, for the preset claim attribute data of "deductible and compensation ratio", the corresponding attribute parameter division intervals include 3 intervals divided based on medical expenses, and these intervals can be flexibly adjusted according to actual needs, which will not be elaborated here.

[0109] In step S320 of some embodiments, further, the present application can adjust the function interval of the claim resource prediction function based on the attribute parameter division intervals to obtain a candidate resource prediction function. At this time, the candidate resource prediction function is the prediction function corresponding to different function intervals.

[0110] In step S330 of some embodiments, further, the present application can substitute the specific value of the target claim attribute data into the candidate resource prediction function to adjust the function parameters and obtain the target resource prediction function of the target object. For example, if the target claim request is a disease claim, and the obtained claim resource prediction function is "predicted claim resource = (number of hospitalization days - deductible days) × compensation ratio × daily resource amount". At this time, for the attribute parameter division interval where the medical expenses are below X1, if the compensation ratio is B1, the deductible days are 3 days, and the daily resource amount is 100, then the target resource prediction function corresponding to this attribute parameter division interval can be "predicted claim resource = (number of hospitalization days - 3) × B1 × 100".

[0111] In the above embodiments, this application can match the target claim attribute data obtained from the target policy text with the corresponding structured attributes, i.e., obtain the attribute parameter division intervals, and then adjust the claim resource prediction function according to the attribute parameter division intervals and the corresponding target claim attribute data. This achieves full automation from data input to claim result output, greatly reducing manual intervention and improving the accuracy and efficiency of claim resource prediction. Furthermore, this application can use a large-scale text parsing model to parse the target policy text, avoiding reliance on manual reading of policy terms and special conditions to determine the applicable claim rules. The method of this application can also adapt to changes in various languages ​​and expressions, improving the accuracy and generalization ability of claim attribute identification.

[0112] See Figure 4 , Figure 4 This is another optional flowchart of the claims resource prediction method provided in the embodiments of this application. In some embodiments of this application, after step S310, the claims resource prediction method provided in the embodiments of this application may further include steps S410 to S440:

[0113] Step S410: Encode the preset claims attribute parameters to obtain the claims parameter codes;

[0114] Step S420: Perform category matching between the target claim attribute data and the preset claim clause categories to obtain the target claim clause categories;

[0115] Step S430: Divide the attribute parameters that match the target claim attribute data into intervals and combine the claim parameter codes into interval codes to obtain structured attribute data;

[0116] Step S440: Combine all structured attribute data that match the target claim terms category to obtain the structured claim data for the target object.

[0117] In step S410 of some embodiments, before obtaining the attribute parameter division interval, this application can first construct structured claims data for the target object, that is, use a structured storage method to store the preset claims attribute parameters in the initial policy text of the target object. Specifically, this application can first encode each preset claims attribute parameter to obtain the corresponding claims parameter code. At this time, the obtained claims parameter code refers to converting text or categorized data into character data that can be processed by machine learning models, enabling the computer system to quickly identify and process it. For example, the preset claims attribute parameters may include accidental medical liability, disease medical liability, resource range, deductible, etc. Then, the claims parameter code corresponding to accidental medical liability can be YWYL, the claims parameter code corresponding to disease medical liability can be JBYL, in the resource range code, QJJED can represent the lowest resource amount in the range, QJJEG can represent the highest resource amount in the range, the claims parameter code corresponding to the deductible can be MPC, and the claims parameter code corresponding to the reimbursement ratio can be PFBL, without specific limitations.

[0118] In step S420 of some embodiments, the target claim attribute data can be further matched with a preset claim clause category to determine which claim clause category the target claim attribute data belongs to. This helps the system understand the specific content of the target claim request and classify it into the correct claim type. At this time, the preset claim clause category may include multiple target claim attribute data, which facilitates improved prediction efficiency in subsequent predictions. For example, the preset claim clause category includes health clauses, vehicle clauses, etc., and the health clause may include different attribute data such as accidental medical liability and illness medical liability.

[0119] In steps S430 and S440 of some embodiments, the attribute parameters matching the target claim attribute data can be further divided into intervals and the claim parameter codes can be combined using interval coding to obtain structured attribute data. All structured attribute data matching the target claim clause category are then combined to obtain complete structured claim data for the target object. This structured claim data contains all the necessary information and can be used for claim decision-making and processing. The purpose of data combination is to create a comprehensive view that allows claims processors to efficiently access and analyze claim information, thereby making accurate claim decisions.

[0120] Interval coding combination refers to dividing continuous data into discrete intervals, which makes data analysis and processing easier. Structured attribute data refers to converting unstructured or semi-structured data into a structured format for easier storage and analysis. For example, when the target claim attribute data is accidental medical liability, the corresponding attribute parameter intervals include {"Minimum resource amount": 0, "Maximum resource amount": 1000, "Deductible": "MPC200", "Payment ratio": "PFBL75"}, {"Minimum resource amount": 1001, "Maximum resource amount": 5000, "Deductible": "MPC300", "Payment ratio": "PFBL80"}, and {"Minimum resource amount": 5001, "Maximum resource amount": null, "Deductible": "MPC400", "Payment ratio": "PFBL85"}.

[0121] In the above embodiments, the structured claims data constructed in this application can be more easily used directly by the claims system for automatic calculation. Each code has a clear meaning, making it easy to identify and process. In addition, this application can simultaneously adopt a multi-level JSON format to make the data clearer and easier to read, effectively improving the accuracy and efficiency of claims resource prediction.

[0122] It should be noted that the large text parsing model used in this application can be self-adjusted and optimized based on actual claims cases, and the structured claims data of this application can be adaptively adjusted according to the modification of the policy text of the target object, thereby ensuring the stability and adaptability of resource prediction performance in the long term.

[0123] In step S180 of some embodiments, finally, the claim resources required for a specific claim request are predicted based on the adjusted target resource prediction function, such as the expected amount of claim resources, the required investigation resources, etc.

[0124] See Figure 5 , Figure 5 This is an optional flowchart of step S180 provided in the embodiments of this application. In some embodiments of this application, step S180 may specifically include steps S510 to S540:

[0125] Step S510: Determine the function prediction conditions of the target resource prediction function based on the target policy text;

[0126] Step S520: Perform data matching on the target claims data based on the prediction condition parameters to obtain claims condition data;

[0127] Step S530: Perform data detection on the claims condition data based on the function prediction conditions to obtain the function state of the target resource prediction function;

[0128] Step S540: Based on the target resource prediction function and the function state, perform claim resource prediction on the target claim data to obtain the target claim resources.

[0129] In step S510 of some embodiments, when performing claims resource prediction on the target claims data, it is also necessary to consider whether the target object meets the conditions for resource prediction, such as insurance period, insurance age requirements, insurance area requirements, etc. The function prediction conditions include prediction condition parameters (e.g., claims resource amount, deductible, insurance period, etc.). The function prediction conditions are used to indicate the conditions that the data in the target claims data that matches the prediction condition parameters must meet when the target resource prediction function is called.

[0130] In step S520 of some embodiments, after determining the prediction condition parameters, this application can match corresponding data from the target claims data, that is, filter and identify data records that match the prediction conditions from the target claims data for subsequent claims resource prediction. For example, the hospitalization duration, examination items, etc. of the target object can be selected from the target object's hospitalization records. In this case, the prediction condition parameters may include a hospitalization duration greater than or equal to a preset duration, such as a preset duration of 10 days.

[0131] In step S530 of some embodiments, the present application may further examine the matched claims condition data to verify whether the data satisfies the function prediction conditions. This examination process includes checks on the integrity, accuracy, and consistency of the data to ensure that the data used for prediction is reliable and valid. Thus, the function state of the target resource prediction function can be determined, which indicates whether the target condition data matched with the prediction condition parameters satisfies the function prediction conditions.

[0132] It should be noted that the function state includes a called state and a non-called state. The called state indicates that the target condition data matching the prediction condition parameters satisfies the function prediction conditions. The non-called state indicates that the target condition data matching the prediction condition parameters does not satisfy the function prediction conditions.

[0133] In step S540 of some embodiments, the present application can further perform claims resource prediction on target claims data based on target resource prediction function and function state, that is, determine whether the target condition data matching the prediction condition parameters meets the function prediction conditions, thereby determining whether to call the corresponding target resource prediction function to perform claims resource prediction.

[0134] See Figure 6 , Figure 6 This is an optional flowchart of step S540 provided in the embodiments of this application. In some embodiments of this application, step S540 may specifically include steps S610 to S620:

[0135] Step S610: If the function is in the calling state, perform data matching on the target claim data based on the preset claim object parameters to obtain the claim object data;

[0136] Step S620: Based on the target resource prediction function and the claim object data, predict the claim resources to obtain the target claim resources.

[0137] In steps S610 and S620 of some embodiments, if the function state is in a calling state, the target resource prediction function can be called. At this time, data matching can be performed on the target claim data based on preset claim object parameters to obtain claim object data. The claim object data is used to characterize the specific data corresponding to the preset claim object parameters in the target claim data. Further, the claim object data can be substituted into the target resource prediction function for calculation to determine the target claim resource. This target claim resource is the predicted resource quantity obtained through automated calculation in this application.

[0138] See Figure 7 , Figure 7 This is another optional flowchart of the claims resource prediction method provided in the embodiments of this application. In some embodiments of this application, after step S180, the claims resource prediction method of this application may further include steps S710 to S740:

[0139] Step S710: Obtain the target object's claim level;

[0140] Step S720: Obtain the resource review model that matches the claim level of the object;

[0141] Step S730: Based on the resource audit model, the target claim resources are audited to obtain the resource status of the target claim resources;

[0142] Step S740: If the resource status indicates that the target claim resource has been approved, transfer the target claim resource.

[0143] In steps S710 and S720 of some embodiments, since the target claim resources predicted by this application may differ from the actual amount of resources paid out by the insurance company, in order to ensure the accuracy of the claim resource prediction, this application may further review the target claim resources to determine whether the predicted target claim resources meet the actual payment requirements. Specifically, this application can match a corresponding resource review model from a preset database according to the claim level of the target object. This resource review model is a set of rules or algorithms used to evaluate and review claim resources. Different claim levels may correspond to different levels of review rigor and procedures. For example, for objects with higher payout ratios or payout limits, the corresponding resource review model is more rigorous, the model complexity is higher, and the review accuracy is higher.

[0144] It should be noted that the claim level of an object can be determined based on the target object's historical claims record, risk assessment, or other relevant factors, and is used to guide the subsequent claims processing procedure.

[0145] In step S730 of some embodiments, the acquired resource review model can be further used to review the target claim resource, determine the resource status of the target claim resource, and the resource status is used to indicate whether the target claim resource has passed or failed the review. The review process may include verifying the completeness, consistency, and compliance of the claim data.

[0146] In step S740 of some embodiments, if the resource status indicates that the target claim resource has been approved, the claim system can continue the subsequent process to realize the resource transfer operation based on the amount of the target claim resource. For example, the approved amount of claim resources can be transferred to the target object's account, or the necessary claim resources can be allocated to meet the target object's claim requirements.

[0147] It should be noted that if the resource status indicates that the target claim resource has failed the review, steps S110 to S180 can be re-executed to redetermine the target claim resource.

[0148] This application provides a method for predicting claims resources, which differs from traditional rule-based text parsing methods. It utilizes a large-scale text parsing model built on a deep learning model, adapting to variations in languages ​​and expressions, thus improving the accuracy and generalization ability of claims attribute identification. Furthermore, this application can directly extract structured attributes from policy text, achieving full automation from data input to claims result output, significantly reducing manual intervention and improving work efficiency. Additionally, the built-in learning framework allows the model to self-adjust and optimize based on actual claims cases, ensuring long-term performance stability and adaptability. Therefore, the claims resource prediction method provided by this application avoids over-reliance on manual processes in the claims process, providing more efficient and accurate claims services, effectively improving the efficiency of claims resource prediction, reducing labor costs, and enabling the claims department to respond more quickly to customer needs. Secondly, the automated claims calculation method reduces the error rate caused by human interpretation, improving the accuracy of the calculation results.

[0149] Please see Figure 8 This application also provides a claims resource prediction device, which can implement the above-mentioned claims resource prediction method. The device includes:

[0150] The first acquisition module 810 is used to acquire the initial policy text of the target object;

[0151] The second acquisition module 820 is used to respond to the target claim request of the target object and acquire the target claim data of the object based on the target claim request;

[0152] The text filtering module 830 is used to filter the initial policy text based on the target claims data to obtain the target policy text.

[0153] The parsing module 840 is used to parse the target policy text based on a preset text parsing model to obtain policy attribute data.

[0154] The third acquisition module 850 is used to acquire a claim resource prediction function based on the target claim request. The claim resource prediction function includes preset claim attribute parameters.

[0155] The data matching module 860 is used to select target claim attribute data that matches the preset claim attribute parameters from the policy attribute data;

[0156] The function adjustment module 870 is used to adjust the claim resource prediction function based on the target claim attribute data to obtain the target resource prediction function for the target object.

[0157] The resource prediction module 880 is used to predict the claim resources of the target claim data based on the target resource prediction function, so as to obtain the target claim resources.

[0158] The specific implementation of the claims resource prediction device is basically the same as the specific implementation of the claims resource prediction method described above, and will not be repeated here.

[0159] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and the processor executes the computer program to implement the aforementioned method for predicting claims resources. This electronic device can be any smart terminal, including tablet computers, in-vehicle computers, etc.

[0160] Please see Figure 9 , Figure 9 The hardware structure of an electronic device according to another embodiment is illustrated. The electronic device includes:

[0161] The processor 910 can be implemented using a general-purpose central processing unit (CPU), microprocessor, application specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of this application.

[0162] The memory 920 can be implemented as a read-only memory (ROM), static storage device, dynamic storage device, or random access memory (RAM). The memory 920 can store the operating system and other applications. When the technical solutions provided in the embodiments of this specification are implemented through software or firmware, the relevant program code is stored in the memory 920 and is called and executed by the processor 910 to execute the claims resource prediction method of the embodiments of this application.

[0163] The input / output interface 930 is used to implement information input and output;

[0164] The communication interface 940 is used to enable communication and interaction between this device and other devices. Communication can be achieved through wired means (such as USB, network cable, etc.) or wireless means (such as mobile network, WIFI, Bluetooth, etc.).

[0165] Bus 950 transmits information between various components of the device (e.g., processor 910, memory 920, input / output interface 930, and communication interface 940);

[0166] The processor 910, memory 920, input / output interface 930 and communication interface 940 are connected to each other within the device via bus 950.

[0167] This application also provides a computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned method for predicting claims resources.

[0168] Memory, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory may optionally include memory remotely located relative to the processor, and these remote memories can be connected to the processor via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0169] The embodiments described in this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided by the embodiments of this application. As those skilled in the art will know, with the evolution of technology and the emergence of new application scenarios, the technical solutions provided by the embodiments of this application are also applicable to similar technical problems.

[0170] Those skilled in the art will understand that the technical solutions shown in the figures do not constitute a limitation on the embodiments of this application, and may include more or fewer steps than shown, or combine certain steps, or different steps.

[0171] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs.

[0172] Those skilled in the art will understand that all or some of the steps in the methods disclosed above, as well as the functional modules / units in the systems and devices, can be implemented as software, firmware, hardware, or suitable combinations thereof.

[0173] The terms “first,” “second,” “third,” “fourth,” etc. (if present) in the specification and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in orders other than those illustrated or described herein. Furthermore, the terms “comprising” and “having,” and any variations thereof, are intended to cover non-exclusive inclusion; for example, a process, method, system, product, or apparatus that comprises a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or apparatus.

[0174] It should be understood that in this application, "at least one (item)" means one or more, and "more than" means two or more. "And / or" is used to describe the relationship between related objects, indicating that three relationships can exist. For example, "A and / or B" can represent three cases: only A exists, only B exists, and both A and B exist simultaneously, where A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one (item) of the following" or similar expressions refer to any combination of these items, including any combination of single or plural items. For example, at least one (item) 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.

[0175] In the several embodiments provided in this application, it should be understood that the disclosed apparatus and methods can be implemented in other ways. For example, the apparatus embodiments described above are merely illustrative; for instance, the division of the units described above is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between apparatuses or units may be electrical, mechanical, or other forms.

[0176] The units described above as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0177] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0178] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes multiple instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of the various embodiments of this application. The aforementioned storage medium includes various media capable of storing programs, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0179] The preferred embodiments of the present application have been described above with reference to the accompanying drawings, but this does not limit the scope of the claims of the present application. Any modifications, equivalent substitutions, and improvements made by those skilled in the art without departing from the scope and substance of the embodiments of the present application shall be within the scope of the claims of the present application.

Claims

1. A method for predicting claims resources, characterized in that, The method includes: Obtain the initial policy text of the target object. The initial policy text refers to the text format of the policy data that the target object has participated in and signed. In response to the target claim request of the target object, and based on the target claim request, obtain the target claim data of the target object; Based on the target claim data, the initial policy text is filtered to obtain the target policy text, which is used to represent the policy text associated with the target claim request. Based on a pre-defined text parsing model, the target policy text is parsed to obtain policy attribute data; A claim resource prediction function is obtained based on the target claim request, and the claim resource prediction function includes preset claim attribute parameters; Select target claim attribute data that matches the preset claim attribute parameters from the policy attribute data; Based on the target claim attribute data, attribute parameter division intervals are obtained from preset structured claim data. These attribute parameter division intervals match the preset claim attribute parameters. The structured claim data is a structured data format stored based on relevant attribute parameters in the initial policy text of the target object, used to quickly query the attribute parameter division intervals corresponding to the target claim attribute data of the target object in the target policy text. Based on the attribute parameter division intervals, the claim resource prediction function is adjusted to obtain candidate resource prediction functions. These candidate resource prediction functions are prediction functions corresponding to different function intervals divided by the attribute parameter division intervals. The specific values ​​of the target claim attribute data are substituted into the candidate resource prediction functions to adjust the function parameters, thus obtaining the target resource prediction function for the target object. Based on the target resource prediction function, the target claims data is used to predict claims resources to obtain target claims resources.

2. The method according to claim 1, characterized in that, The step of predicting the target claims resources based on the target claims data using the target resource prediction function to obtain the target claims resources includes: Based on the target policy text, the function prediction conditions of the target resource prediction function are determined. The function prediction conditions include prediction condition parameters. The function prediction conditions are used to indicate the conditions that the data in the target claims data that matches the prediction condition parameters must meet when the target resource prediction function is called. Based on the predicted condition parameters, the target claim data is matched to obtain claim condition data; Based on the function prediction conditions, data detection is performed on the claims condition data to obtain the function state of the target resource prediction function. The function state is used to indicate that the target condition data that matches the prediction condition parameters satisfies the function prediction conditions. Based on the target resource prediction function and the function state, the target claim data is used to predict the claim resources, thereby obtaining the target claim resources.

3. The method according to claim 2, characterized in that, The function state includes a call state, which is used to indicate that the target condition data that matches the prediction condition parameters satisfies the function prediction conditions. The claims resource prediction function also includes preset claims object parameters. The step of predicting claim resources based on the target claim data using the target resource prediction function and the function state to obtain the target claim resources includes: If the function is in the calling state, the target claim data is matched based on the preset claim object parameters to obtain the claim object data; Based on the target resource prediction function and the claims object data, the claims resource is predicted to obtain the target claims resource.

4. The method according to claim 1, characterized in that, After obtaining the attribute parameter division range from the preset structured claims data based on the target claims attribute data, the method further includes: The preset claims attribute parameters are encoded to obtain claims parameter codes; The target claim attribute data and the preset claim clause categories are matched to obtain the target claim clause category; The attribute parameters that match the target claim attribute data are divided into intervals and the claim parameter codes are combined using interval coding to obtain structured attribute data; The structured attribute data that matches the target claim clause category are combined to obtain the structured claim data for the target object.

5. The method according to any one of claims 1 to 4, characterized in that, The process of obtaining the initial policy text of the target object includes: Obtain the object account information of the target object; Based on the object's account information, the original policy data is matched from a preset database; The original policy data is subjected to text recognition to obtain the original policy text; The original policy text is preprocessed to obtain the initial policy text.

6. The method according to any one of claims 1 to 4, characterized in that, The method further includes: Obtain the object claim level of the target object; Obtain the resource review model that matches the claim level of the object; Based on the resource review model, the target claim resources are reviewed to obtain the resource status of the target claim resources. The resource status is used to indicate whether the target claim resources have passed the review or failed the review. If the resource status indicates that the target claim resource has been approved, the target claim resource will be transferred.

7. A device for predicting claims resources, characterized in that, The device includes: The first acquisition module is used to acquire the initial policy text of the target object. The initial policy text refers to the text form of the policy data that the target object has participated in and signed. The second acquisition module is used to respond to the target claim request of the target object and acquire the target claim data of the object based on the target claim request; The text filtering module is used to filter the initial policy text based on the target claim data to obtain the target policy text, which is used to represent the policy text associated with the target claim request. The parsing module is used to parse the target policy text based on a preset text parsing model to obtain policy attribute data; The third acquisition module is used to acquire a claim resource prediction function based on the target claim request, wherein the claim resource prediction function includes preset claim attribute parameters; The data matching module is used to select target claim attribute data that matches the preset claim attribute parameters from the policy attribute data; The function adjustment module is used to obtain attribute parameter division intervals from preset structured claims data based on the target claims attribute data. The attribute parameter division intervals match the preset claims attribute parameters. The structured claims data is a structured data format stored based on relevant attribute parameters in the initial policy text of the target object. It is used to quickly query the attribute parameter division intervals corresponding to the target claims attribute data of the target object in the target policy text. Based on the attribute parameter division intervals, the module adjusts the function intervals of the claims resource prediction function to obtain candidate resource prediction functions. The candidate resource prediction functions are prediction functions corresponding to different function intervals divided by the attribute parameter division intervals. The module substitutes the specific values ​​of the target claims attribute data into the candidate resource prediction functions to adjust the function parameters, thereby obtaining the target resource prediction function for the target object. The resource prediction module is used to predict the claim resources of the target claim data based on the target resource prediction function, so as to obtain the target claim resources.

8. An electronic device, characterized in that, The electronic device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the method according to any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by a processor, it implements the method of any one of claims 1 to 6.

Images