Data processing method and device, equipment, storage medium and product
By extracting key elements from user-submitted claim information and using accident environment information to verify and revise claim plans, the problem of low efficiency in manual processing in existing technologies has been solved, achieving more efficient and accurate claim processing.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CHINA MOBILE FINANCIAL TECHNOLOGY CO LTD
- Filing Date
- 2025-01-06
- Publication Date
- 2026-07-07
Smart Images
Figure CN122347474A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of data processing technology, and in particular to data processing methods, apparatus, equipment, storage media and products. Background Technology
[0002] Property insurance refers to insurance where the insured pays a premium to the insurer according to the contract, and the insurer assumes liability for compensation for losses to the insured property and related interests caused by natural disasters or accidents, as stipulated in the insurance contract. Property loss insurance mainly uses various tangible properties as the insured object. In related technology, after a property insurance claim occurs, the insured reports the incident and provides details of the loss, then submits relevant documents to apply for a claim. The insurance company conducts a manual investigation and verification, reviews the materials and assesses the loss, and finally pays compensation. However, the efficiency of manually handling property insurance claims is relatively low, especially when facing a large number of cases or major insurance events, where the problems of manpower shortage and limited claims processing capacity become more prominent. Summary of the Invention
[0003] The main objective of this application is to provide a data processing method, apparatus, device, storage medium, and product, which aims to solve the technical problem of low efficiency in the existing manual processing of property insurance claims.
[0004] To achieve the above objectives, this application proposes a data processing method, which includes:
[0005] Extract key claim elements from the claim information submitted by the user, and determine the initial claim settlement plan based on the key claim elements;
[0006] Obtain the accident environment information corresponding to the claim information, verify the key elements of the claim based on the accident environment information, and obtain the verification result;
[0007] If the verification result is successful, the initial claim settlement plan is modified based on the accident environment information to obtain the target claim settlement plan.
[0008] Optionally, after the step of revising the initial claim settlement plan based on the accident environment information to obtain the target claim settlement plan when the verification result is successful, the method further includes:
[0009] Determine the confidence level of the target claims settlement plan;
[0010] If the confidence level meets the preset confidence level conditions, the accident information is verified through a smart contract, and compensation is paid to the user after the verification is successful.
[0011] Optionally, the step of extracting key claim elements from the claim information submitted by the user and determining the initial claim settlement plan based on the key claim elements includes:
[0012] The user-submitted claim information is converted into text format to obtain the claim information.
[0013] Entity recognition is performed based on the claim information in the text format to obtain the accident entity recognition result;
[0014] The key elements of the claim are determined based on the accident entity identification results and the claim information in the text format.
[0015] The initial claims settlement plan is determined based on the aforementioned key elements of claims settlement and the preset compensation model.
[0016] Optionally, before the step of determining the initial claims settlement plan based on the key claims elements and the preset compensation model, the method further includes:
[0017] Obtain policy recall information, and determine sample data and policy recall rate based on the policy recall information;
[0018] The initial compensation model is trained based on the sample data and the policy recall rate to obtain the preset compensation model.
[0019] Optionally, the step of verifying the key elements of the claim based on the accident environment information to obtain the verification result includes:
[0020] Feature extraction is performed on the accident environment information to obtain accident environment features;
[0021] Determine the accident environment feature vector based on the aforementioned accident environment characteristics;
[0022] Determine the feature vector of the key claim elements based on the aforementioned key claim elements;
[0023] Determine the degree of matching between the accident environment feature vector and the feature vector of the key elements of the claims settlement;
[0024] The verification results of the key elements of the claim are determined based on the matching degree.
[0025] Optionally, the step of revising the initial claim settlement plan based on the accident environment information to obtain the target claim settlement plan when the verification result is successful includes:
[0026] If the verification result is successful, the driver's degree of responsibility is determined based on the accident environment information;
[0027] The initial claim settlement plan is modified based on the liability ratio to obtain the target claim settlement plan.
[0028] Furthermore, to achieve the above objectives, this application also proposes a data processing apparatus, the data processing apparatus comprising:
[0029] The initial claims settlement plan determination module is used to extract key claims elements from the claims information submitted by the user and determine the initial claims settlement plan based on the key claims elements.
[0030] The verification module is used to obtain the accident environment information corresponding to the claim information, verify the key elements of the claim based on the accident environment information, and obtain the verification result.
[0031] The initial claims settlement plan modification module is used to modify the initial claims settlement plan based on the accident environment information when the verification result is successful, so as to obtain the target claims settlement plan.
[0032] In addition, to achieve the above objectives, this application also proposes a data processing apparatus, the apparatus comprising: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the data processing method described above.
[0033] In addition, to achieve the above objectives, this application also proposes a storage medium, which is a computer-readable storage medium, on which a computer program is stored, and which, when executed by a processor, implements the steps of the data processing method described above.
[0034] In addition, to achieve the above objectives, this application also provides a computer program product, which includes a computer program that, when executed by a processor, implements the steps of the data processing method described above.
[0035] This application extracts key claim elements from user-submitted claim information and determines an initial claim settlement plan based on these elements. It then obtains accident environment information corresponding to the claim information and verifies the key claim elements based on this information, obtaining a verification result. If the verification result is successful, the initial claim settlement plan is revised based on the accident environment information to obtain a target claim settlement plan. Because this application verifies key claim elements in user-submitted claim information based on accident environment information, and then revises the initial claim settlement plan based on the accident environment information to obtain a target claim settlement plan, compared to existing methods of manually reviewing and evaluating user-submitted claim information, this method improves claim settlement efficiency and accuracy. Attached Figure Description
[0036] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application.
[0037] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, for those skilled in the art, other drawings can be obtained based on these drawings without creative effort.
[0038] Figure 1 This is a flowchart illustrating an embodiment of the data processing method of this application.
[0039] Figure 2 This is a flowchart illustrating Embodiment 2 of the data processing method of this application;
[0040] Figure 3 This is a schematic diagram of entity relationships provided in Embodiment 2 of the data processing method of this application;
[0041] Figure 4 This is a schematic diagram of the overall process of the data processing method in Embodiment 2 of this application;
[0042] Figure 5 This is a schematic diagram of the module structure of the data processing device according to an embodiment of this application;
[0043] Figure 6 This is a schematic diagram of the device structure of the hardware operating environment involved in the data processing method in the embodiments of this application.
[0044] The purpose, features, and advantages of this application will be further explained in conjunction with the embodiments and with reference to the accompanying drawings. Detailed Implementation
[0045] It should be understood that the specific embodiments described herein are merely illustrative of the technical solutions of this application and are not intended to limit this application.
[0046] To better understand the technical solution of this application, a detailed description will be provided below in conjunction with the accompanying drawings and specific implementation methods.
[0047] The main solution of this application is as follows: extract key claim elements from the claim information submitted by the user, determine an initial claim plan based on the key claim elements; obtain accident environment information corresponding to the claim information, verify the key claim elements based on the accident environment information, and obtain a verification result; if the verification result is successful, modify the initial claim plan based on the accident environment information to obtain a target claim plan. Since this application verifies the key claim elements in the claim information submitted by the user based on the accident environment information, and modifies the initial claim plan based on the accident environment information to obtain a target claim plan when the verification is successful, compared to the existing method of manually reviewing and evaluating the claim information submitted by the user, the above method of this application can improve claim efficiency and accuracy.
[0048] It should be noted that the executing entity in this embodiment can be a computing service device with data processing, network communication, and program execution functions, such as a tablet computer, personal computer, or mobile phone, or an electronic device or data processing device capable of performing the above functions. The following description uses a data processing device as an example to illustrate this embodiment and the subsequent embodiments.
[0049] Based on this, embodiments of this application provide a data processing method, referring to... Figure 1 , Figure 1 This is a flowchart illustrating an embodiment of the data processing method of this application.
[0050] In this embodiment, the data processing method includes the following steps:
[0051] Step S10: Extract key claim elements from the claim information submitted by the user, and determine the initial claim settlement plan based on the key claim elements;
[0052] It should be noted that the claims information may include user-submitted claims applications, claims evidence, and third-party assessment reports. The types of claims information include structured data (e.g., tables in a database), semi-structured data (e.g., XML, JSON format), or unstructured data (e.g., text, images, or videos). The key claims elements refer to information points crucial to the claims decision, including the accident type, lost items, extent of loss, time of accident, and location of accident. Determining the initial claims plan based on the key claims elements may involve inputting the key claims elements into a preset compensation model to obtain the claims plan output by the preset compensation model. The initial claims plan may include the amount of compensation. The preset compensation model may be a model pre-trained on a neural network model using sample data, capable of predicting the initial claims plan based on information such as key claims elements.
[0053] Step S20: Obtain the accident environment information corresponding to the claim information, verify the key elements of the claim based on the accident environment information, and obtain the verification result;
[0054] It should be noted that the accident environment information can be image and video data of the accident scene collected based on the time and location of the property accident. This data can come from vehicle dashcams, surveillance cameras, drones, smartphones, etc. The verification of the key elements of the claim based on the accident environment information, and the resulting verification result, can be a determination of the authenticity of the information in the key elements of the claim based on the accident environment information.
[0055] Furthermore, to prevent users from submitting false claim information, step S20 includes: extracting features from the accident environment information to obtain accident environment features;
[0056] Determine the accident environment feature vector based on the aforementioned accident environment characteristics;
[0057] Determine the feature vector of the key claim elements based on the aforementioned key claim elements;
[0058] Determine the degree of matching between the accident environment feature vector and the feature vector of the key elements of the claims settlement;
[0059] The verification results of the key elements of the claim are determined based on the matching degree.
[0060] It should be noted that determining the accident environment feature vector based on the accident environment characteristics can be achieved by vectorizing the accident environment characteristics to obtain the accident environment feature vector. Similarly, determining the feature vector of the key claim elements based on the key claim elements is done in the same way. The degree of matching between the accident environment feature vector and the feature vector of the key claim elements can be determined by the following formula:
[0061]
[0062] Where M represents the matching degree between the accident environment feature vector and the claim key element feature vector, E1 represents the accident environment feature vector, and E0 represents the claim key element feature vector. The closer M is to 1, the more similar the basic accident environment information features are to the claim key elements, and the lower the possibility of user fraud.
[0063] It should be noted that the verification result of determining the key elements of the claim based on the matching degree can be determined as verification passed when the matching degree is less than a preset threshold.
[0064] In practice, accident environment features corresponding to key claim elements are extracted from the accident environment information to verify the correctness of the key claim elements. Specifically, based on the key claim elements in the entity relationship graph, accident environment information is retrieved in reverse to verify the correctness of the key claim elements. For example, based on the time and location of the property accident, images and video data of the accident scene are collected. This data can come from various sources such as vehicle dashcams, surveillance cameras, drones, and smartphones. Feature extraction algorithms in computer vision technology are used to extract basic accident environment information features (such as the time, location, and people involved in the accident) from the images or video data of the accident scene. The extracted basic accident environment information features are then matched with the key claim elements. Extracting basic accident environment information features from the images or video data of the accident scene includes the following steps:
[0065] a. Perform preprocessing operations such as denoising and contrast enhancement on images or video frames from the accident scene to improve the accuracy of subsequent processing. If processing video data, keyframe extraction is required first to reduce computational load and retain important information.
[0066] b. Use the Scale-Invariant Feature Transform (SIFT) algorithm to extract features from the preprocessed image or keyframes, obtaining local invariant feature points and their feature descriptors. This includes the following steps:
[0067] b1. Scale-space extremum detection and key point localization: A scale space is generated through Gaussian blur and difference operations, and local extremum points are searched at different scales as key points.
[0068] b2. Direction Assignment: Select a neighborhood around the keypoint and calculate the gradient direction and magnitude of the pixels within that neighborhood. Use a histogram to statistically analyze the gradient directions of the pixels surrounding the keypoint, and take the peak direction of the histogram as the dominant direction of the keypoint. A keypoint may have multiple dominant directions, and each direction corresponds to a feature descriptor.
[0069] b3. Feature Descriptor Generation: In the local image region surrounding the key point, the gradient direction and magnitude are calculated. This gradient information is encoded into a descriptor vector with a fixed length, which is robust to various image transformations to accurately match the same feature points in different images. After determining the feature descriptors, the basic environmental information features of the accident are determined by comparing the feature descriptors at different times. Among them: Temporal features: The approximate time period of the accident (e.g., daytime, nighttime) is indirectly inferred by analyzing the lighting conditions and shadow direction in the image. Alternatively, the time stamp in the video data is combined to determine the event of the accident.
[0070] Location Features: By analyzing background information in images (such as buildings, road signs, and natural landscapes) and combining it with Geographic Information System (GIS) data, the location of the accident is identified. The location identification result is associated with the corresponding SIFT descriptors as location features. This is achieved by creating a data structure (such as a dictionary, list, or database record) where the keys are location labels and the values are a list of SIFT descriptors corresponding to that location.
[0071] Person Features: Person features include facial features and pose features. Deep learning models (such as convolutional neural networks) are used to detect faces in images and extract facial features. Pose estimation models are used to detect human bodies in images and extract human pose features. The person recognition results are associated with the corresponding SIFT descriptors as person features. This is achieved by creating a data structure where the keys are person feature vectors and the values are lists of SIFT descriptors corresponding to that person.
[0072] Accident Type Features: Combining vehicle and pedestrian detection results with clues such as collision marks and scattered objects in the image, the accident type (e.g., rear-end collision, rollover, pedestrian collision, etc.) is analyzed. The accident type identification result is associated with the corresponding SIFT descriptor as the accident type feature. Once the accident type is identified, key regions related to the accident type need to be located in the image. The SIFT algorithm is applied to these key regions to extract feature points and their descriptors, implemented by creating a data structure (e.g., a dictionary), where the key is the accident type label and the value is a list of SIFT descriptors corresponding to that accident type. The basic accident environmental information features and key claim elements are transformed into a basic accident environmental information feature vector (i.e., the accident environmental feature vector) and a feature vector of the key claim elements, respectively. The extracted and encoded basic accident environmental information features are combined into a feature vector. Each element in this vector corresponds to a feature, and its value represents the specific value or encoded result of that feature. First, it is determined which basic accident environmental information is used as features. Then, each feature is encoded, and after encoding all features, a feature vector is constructed. Each element of the feature vector corresponds to a feature, and its value is the encoded result of that feature. The length of the feature vector is equal to the number of all selected features. For example, if the year is 2024 (using the numerical value 2024 directly), the month is September (using one-hot encoding [0, 0, 0, 1, 0, 0, 0, 0, 0, 1, 0, 0] to represent January to December), the weather is sunny (one-hot encoding [1, 0, 0] to represent sunny, rainy, and snowy), and the road condition is clear (one-hot encoding [1, 0, 0] to represent clear, congested, and under construction), then the constructed basic environmental information feature vector for the accident is: Basic Environmental Information Feature Vector for Accident = [2024, 0, 0, 0, 1, 0, 0, 0, 0, 1, 0, 0, 1, 0, 0, 1, 0, 0]. This basic environmental information feature vector for the accident contains information from all selected features, and each feature has been encoded into a format that the model can process. Similarly, the extracted and encoded key claim element features are combined into a feature vector. The structure of this vector is similar to that of the basic accident environment information feature vector. Then, the matching degree between the accident environment feature vector and the key claim element feature vector is determined, and the verification result of the key claim element is determined based on the matching degree.
[0073] Step S30: If the verification result is successful, the initial claim settlement plan is modified based on the accident environment information to obtain the target claim settlement plan.
[0074] It should be noted that the step of modifying the initial claim settlement plan based on the accident environment information to obtain the target claim settlement plan can be achieved by determining the user's liability ratio based on the accident environment information, and then modifying the compensation amount in the initial claim settlement plan according to the liability ratio to obtain the target claim settlement plan.
[0075] Furthermore, in order to revise the initial claim settlement plan according to the user's proportion of responsibility, step S30 may include: if the verification result is that the verification is passed, determining the driver's proportion of responsibility based on the accident environment information;
[0076] The initial claim settlement plan is modified based on the liability ratio to obtain the target claim settlement plan.
[0077] It should be noted that this embodiment can utilize deep learning models to comprehensively analyze the environmental information of the accident scene, including image recognition technology to identify objects and traces at the accident scene, and text analysis technology to process unstructured data such as accident reports and witness statements. Environmental information features of the accident scene are extracted for subsequent determination of liability proportions. After determining the liability proportions of the responsible parties, the compensation amount in the initial claims settlement plan is adjusted based on these proportions. For example, for automobile property insurance, accident environmental information features include driver behavior characteristics, weather conditions, road conditions, and traffic conditions. For mobile phone property insurance, accident environmental information features include user behavior characteristics and weather conditions. Taking automobile property insurance as an example, the determination of liability proportions can be achieved by standardizing the collected data of each factor to ensure that the data is compared on the same scale. Driver behavior characteristics (such as speeding, illegal lane changes, etc.), weather conditions (such as rain, snow, haze, etc.), road conditions (such as slippery, potholes, under maintenance, etc.), and traffic conditions (such as traffic flow, traffic lights, etc.) are quantified and scored according to their impact on accidents.
[0078] The proportion of responsibility shall be determined according to the following formula:
[0079]
[0080] Where 'a' represents the liability ratio of the party whose property was damaged, β1, β2, β3, and β4 are the weights of driver behavior characteristics, weather conditions, road conditions, and traffic conditions, respectively, β is the total weight, and G1, G2, G3, and G4 are the scores of driver behavior characteristics, weather conditions, road conditions, and traffic conditions, respectively, β = β1 + β2 + β3 + β4. If the driver's behavior characteristic is illegal lane changing, but the road was slippery that day, introducing road conditions into the determination of liability ratio will weaken the driver's illegal behavior, reduce the driver's subjective responsibility, and improve the user's insurance claim experience. The compensation amount can be adjusted as follows: after determining the liability ratio of the party responsible for the accident, the compensation amount in the initial claim plan can be adjusted based on this ratio. For example, if the initial compensation amount is A and the liability ratio is B (expressed as a percentage, such as 50% is 0.5), then the adjusted compensation amount C can be calculated using the following formula: C = A * B.
[0081] Furthermore, in order to reduce claims errors, after step S30, the method further includes: determining the confidence level of the target claims settlement plan;
[0082] If the confidence level meets the preset confidence level conditions, the accident information is verified through a smart contract, and compensation is paid to the user after the verification is successful.
[0083] It should be noted that confidence level can be understood as the degree of certainty regarding the target claim settlement plan, i.e., the probability that the model considers the target claim settlement plan correct. Determining the confidence level of the target claim settlement plan can be achieved through a measurement model. Specifically, the calculation process for the confidence level of the claim settlement plan is as follows: the output of the softmax function is used as the probability distribution of the predicted category, where the maximum probability value is used as an approximate estimate of the confidence level. A neural network model is specifically chosen as the measurement model, which can evaluate the reliability of the claim settlement decision plan based on key claim elements, compensation rules, and accident environment information. The output layer of the measurement model uses the softmax activation function, which transforms the output z of the measurement model into a probability distribution p, where each element p... i This represents the probability of element i.
[0084] The formula for the softmax function is as follows:
[0085]
[0086] Where n is the total number of elements in the output vector z, zi is the i-th element of the output vector z, and p i It is the probability that it is i.
[0087] After obtaining the probability distribution p, the value with the highest probability is chosen as an approximate estimate of the confidence level. That is:
[0088] U = max(p1, p2, ..., p) n )
[0089] Where U is the confidence level, p1 is the probability distribution of the first element, p2 is the probability distribution of the second element, and p... n Let be the probability distribution of the nth element.
[0090] It should be noted that the pre-set confidence conditions may include a pre-set confidence threshold. When the confidence level of the target claim settlement plan is greater than the pre-set confidence threshold, the pre-set confidence conditions are deemed met. At this point, the accident information is verified through a smart contract on the blockchain, and compensation is paid to the user upon successful verification. Specifically, a smart contract is deployed on the blockchain network. The smart contract contains the logic of the claim settlement decision and the relevant rules for paying compensation. When the confidence level meets the pre-set confidence conditions, the smart contract automatically receives accident information from the claim settlement decision system. The accident information includes key claim elements (such as accident type, list of lost items), accident environment information (such as the time and location of the accident), and the proportion of responsibility. Then, the smart contract verifies the accident information, including:
[0091] (1) Internal information verification: Check whether the incident information conforms to the established business rules and logical constraints. For example, verify whether the incident type is within the scope supported by the contract, and whether the items in the list of lost items match the actual items.
[0092] (2) External data source interaction verification: Interact and verify with data sources on other blockchains. For example, verify whether the weather conditions at the time of the accident are consistent with the accident description by calling the weather information contract on the blockchain; or confirm whether the accident vehicle is insured by querying the vehicle insurance information contract on the blockchain, etc.
[0093] (3) Signature Verification: In a blockchain environment, accident information is often digitally signed by relevant parties (such as the parties involved in the accident, insurance companies, etc.). Smart contracts verify these signatures to ensure the authenticity of the information and that it has not been tampered with.
[0094] This embodiment extracts key claim elements from the claim information submitted by the user, determines an initial claim settlement plan based on these elements, obtains the accident environment information corresponding to the claim information, verifies the key claim elements based on the accident environment information, and obtains a verification result. If the verification result is successful, the initial claim settlement plan is modified based on the accident environment information to obtain a target claim settlement plan. Since this embodiment verifies the key claim elements in the user-submitted claim information based on the accident environment information, and modifies the initial claim settlement plan based on the accident environment information to obtain a target claim settlement plan, compared to the existing method of manually reviewing and evaluating user-submitted claim information, this embodiment can improve claim settlement efficiency and accuracy.
[0095] Based on the first embodiment of this application, in the second embodiment of this application, the content that is the same as or similar to that in the first embodiment described above can be referred to the above description, and will not be repeated hereafter. Based on this, please refer to... Figure 2 , Figure 2 This is a flowchart illustrating a second embodiment of the data processing method of this application. Step S10 further includes the following steps:
[0096] Step S101: Convert the claim information submitted by the user into text to obtain claim information in text format;
[0097] It should be noted that the types of claim information submitted by users include structured data (such as tables in a database), semi-structured data (such as XML and JSON formats), and unstructured data (such as text, images, or videos). The claim information needs to be converted into text format. Specifically, this may include cleaning unstructured claim text data, removing irrelevant characters and punctuation, and performing word segmentation and recognition on accident information entities to divide the text into entities related to the accident information.
[0098] Step S102: Perform entity recognition based on the claim information in the text format to obtain the accident entity recognition result;
[0099] It should be noted that the entity recognition based on the text-formatted claims information can be to identify entities in the text-formatted claims information, such as location, time, and people.
[0100] Step S103: Determine the key elements of the claim based on the accident entity identification results and the text-formatted claim information;
[0101] It should be noted that determining the key elements of the claim based on the accident entity identification results and the text-formatted claim information may involve extracting the relationships between entities in the accident entity identification results from the text-formatted claim information, constructing an entity relationship graph based on the entities and the extracted relationships, and determining the key elements of the claim based on the entity relationship graph.
[0102] In practical implementation, the converted claim text can be processed using Natural Language Processing (NLP) technology to extract key claim elements (including entities and their relationships) from the claim data and represent them as an entity relationship diagram. The specific implementation method is as follows:
[0103] 1. Word segmentation and recognition of accident information entities. Unstructured claims text data is cleaned to remove irrelevant characters and punctuation, and word segmentation and recognition of accident information entities are performed to divide the text into entities related to the accident information. For example, the transformed claims text is: "On March 15, 2024, Zhang San and Li Si were involved in a traffic accident on Huanghe Road, resulting in serious damage to the vehicle. The vehicle owner is Zhang San, and Zhang San and Li Si are a married couple." The identified accident information entities are: "March 15, 2024", "Zhang San", "Li Si", "Huanghe Road", "traffic accident", and "vehicle".
[0104] 2. Extract the relationships between accident information entities. Following the logical order of accident occurrence time, property owner, parties involved in the accident, accident location, accident type, and lost items, extract the relationships between accident information entities. For example, the extracted relationships could be: "The accident occurred on March 15, 2024," "The property owner is Zhang San," "The parties involved in the accident are Zhang San and Li Si," "The accident location is Huanghe Road," "The accident type is a traffic accident," and "The lost items are vehicles."
[0105] 3. Construct an entity-relationship diagram. Represent accident information entities and their relationships graphically. When constructing the entity-relationship diagram, it's necessary not only to extract relationships between entities but also relationships between entities and detailed information (obtainable by retrieving user-entered information). For example, extract the relationship between "vehicle" and "vehicle model" and include this relationship in the entity-relationship diagram. Use a regular expression R to match detailed information in a specific format (e.g., vehicle model). Treat the regular expression R as a function or mapping, mapping a text string T to one or more matching substring sets M. R:T->M, where R is the regular expression, M is the set of all substrings matching the vehicle model, and T is the text string. If there is no substring in T that matches R, then M is an empty set. Then, construct the entity-relationship diagram based on the entities and extracted relationships, and determine the key elements for claims. The key elements for claims can be the textual representation of the entity-relationship diagram. The constructed entity-relationship diagram can be referenced... Figure 3 , Figure 3 This is a schematic diagram of entity relationships provided for Embodiment 2 of the data processing method of this application.
[0106] Step S104: Determine the initial claims settlement plan based on the key elements of the claims settlement and the preset compensation model.
[0107] It should be noted that determining the initial claim settlement plan based on the key claim elements and the preset compensation model can be achieved by inputting the key claim elements and compensation rules into the preset compensation model to obtain the initial claim settlement plan output by the preset compensation model.
[0108] Furthermore, in order to accurately determine the claims settlement plan, before step S104, the following steps may be included: obtaining policy recall information, and determining sample data and policy recall rate based on the policy recall information;
[0109] The initial compensation model is trained based on the sample data and the policy recall rate to obtain the preset compensation model.
[0110] It should be noted that in this embodiment, relevant data on policy recall rate need to be continuously collected and monitored, including the reasons for recall, time, types of policies involved, and compensation amount, in order to iteratively train the initial compensation model and make the claims settlement plan predicted by the obtained preset compensation model more accurate.
[0111] In practice, when the policy recall rate exceeds a set value, the fastest-recalling policies are collected and labeled as negative samples, while policies not recalled within a preset period (e.g., the previous year) are collected and labeled as positive samples. The positive and negative samples are then divided into training, validation, and test sets according to a set ratio (e.g., 70%, 15%, 15%). The initial compensation model is trained using the training, validation, and test sets, and its parameters are modified. These parameters include the regularization coefficient and the learning rate. The learning rate refers to the step size for updating model parameters during machine learning or deep learning. The regularization coefficient is a parameter used in machine learning to impose a constraint (penalty) on the objective function. The regularization coefficient is adjusted based on the policy recall rate: if the recall rate is low, it may be due to an overly simple or underfitting model. In this case, the regularization coefficient can be reduced to allow for greater flexibility in the model parameters, thereby increasing the model's fitting ability. Conversely, if the recall is too high but the precision is low (i.e., the model identifies many positive samples, but many of them are incorrect), it may be due to the model being too complex or overfitting. In this case, you can try increasing the regularization coefficient to reduce the model's complexity. Adjust the learning rate based on the policy recall: If the recall fluctuates greatly, you can try decreasing the adjustment learning rate. A smaller learning rate can make the model training more stable, but it may increase training time. Conversely, if the model converges slowly, you can try increasing the adjustment learning rate. A larger learning rate can speed up the training process, but it may lead to instability in the training process.
[0112] In specific implementation, it can be referred to Figure 4 , Figure 4 This is a schematic diagram of the overall process for the data processing method embodiment two of this application. This embodiment aims to achieve accurate automatic claims settlement for property insurance, and mainly includes six key steps: extracting key elements of the claims settlement; verifying key elements of the claims settlement based on accident environment information; analyzing accident environment information to determine the proportion of liability, adjusting the compensation amount, and obtaining the target claims settlement plan; determining the confidence level of the target claims settlement plan; verifying accident information via smart contract; and adjusting the compensation model parameters based on the policy recall rate. Specifically: S1, analyze various types of claims data from property insurance, convert them into claims text, and extract key elements of the claims settlement from the claims text. The claims data includes customer-submitted claims applications, claims evidence, and third-party assessment reports. The types of claims data include structured data (e.g., tables in a database), semi-structured data (e.g., XML, JSON format), or unstructured data (e.g., text, images, or videos). Key elements of the claims settlement refer to information points crucial to the claims settlement decision, including the type of accident, the items lost, the extent of the loss, the time of the accident, and the location of the accident.
[0113] S2. Extract the accident environment information corresponding to the key elements of the claim from the accident environment information, and verify the correctness of the key elements of the claim.
[0114] S3. After verifying the correctness of the key claim elements, combine the key claim elements and compensation rules, and use the compensation model to provide a property insurance claim decision plan. The compensation rules are defined according to the insurance company's compensation policies and regulations, covering the correspondence between different accident types, loss levels, and compensation amounts, as well as other factors that may affect the compensation amount (such as policy terms, deductibles, etc.). The compensation model can be built using deep learning technology. Inputting the integrated key claim elements into the compensation model, the model will reason and calculate according to the compensation rules, generating a preliminary claim decision plan, including the expected compensation amount, a recommendation on whether the claim should be approved, and possible claim conditions or limitations.
[0115] S4. Analyze the environmental information of the accident, determine the proportion of responsibility of the responsible party, and revise the compensation amount in the claims decision plan.
[0116] S5. Determine the confidence level of the revised claims decision plan.
[0117] S6. If the confidence level meets the requirements, use smart contracts on blockchain technology to verify the accident information and automatically pay compensation to the user.
[0118] S7. Adjust the model parameters of the compensation model based on the policy recall rate in the later stages.
[0119] This embodiment converts user-submitted claim information into text-formatted claim information; performs entity recognition based on the text-formatted claim information to obtain accident entity recognition results; determines key claim elements based on the accident entity recognition results and the text-formatted claim information; and determines an initial claim settlement plan based on the key claim elements and a preset compensation model. This embodiment extracts key claim elements crucial to claim decisions from various types of claim data submitted by users, improving the efficiency and accuracy of claims processing through automated extraction of key claim elements.
[0120] It should be noted that the above examples are only for understanding this application and do not constitute a limitation on the data processing method of this application. Any simple modifications based on this technical concept are within the protection scope of this application.
[0121] This application also provides a data processing apparatus, please refer to... Figure 5 The data processing device includes:
[0122] The initial claim settlement plan determination module 10 is used to extract key claim settlement elements from the claim information submitted by the user and determine the initial claim settlement plan based on the key claim settlement elements.
[0123] Verification module 20 is used to obtain accident environment information corresponding to the claim information, verify the key elements of the claim based on the accident environment information, and obtain verification results;
[0124] The initial claim plan correction module 30 is used to correct the initial claim plan based on the accident environment information when the verification result is successful, so as to obtain the target claim plan.
[0125] This embodiment extracts key claim elements from the claim information submitted by the user, determines an initial claim settlement plan based on these elements, obtains the accident environment information corresponding to the claim information, verifies the key claim elements based on the accident environment information, and obtains a verification result. If the verification result is successful, the initial claim settlement plan is modified based on the accident environment information to obtain a target claim settlement plan. Since this embodiment verifies the key claim elements in the user-submitted claim information based on the accident environment information, and modifies the initial claim settlement plan based on the accident environment information to obtain a target claim settlement plan, compared to the existing method of manually reviewing and evaluating user-submitted claim information, this embodiment can improve claim settlement efficiency and accuracy.
[0126] The data processing apparatus provided in this application, employing the data processing method described in the above embodiments, can solve the technical problem of low efficiency in the existing manual processing of property insurance claims. Compared with the prior art, the beneficial effects of the data processing apparatus provided in this application are the same as those of the data processing method provided in the above embodiments, and other technical features in the data processing apparatus are the same as those disclosed in the methods of the above embodiments, and will not be repeated here.
[0127] This application provides a data processing apparatus, which includes: at least one processor; and a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, and the instructions are executed by the at least one processor to enable the at least one processor to perform the data processing method in Embodiment 1 above.
[0128] The following is for reference. Figure 6This document illustrates a structural diagram of a data processing device suitable for implementing embodiments of this application. The data processing device in these embodiments may include, but is not limited to, mobile terminals such as mobile phones, laptops, digital broadcast receivers, PDAs (Personal Digital Assistants), PADs (Portable Application Description), PMPs (Portable Media Players), and in-vehicle terminals (e.g., in-vehicle navigation terminals), as well as fixed terminals such as digital TVs and desktop computers. Figure 6 The data processing device shown is merely an example and should not impose any limitations on the functionality and scope of use of the embodiments of this application.
[0129] like Figure 6 As shown, the data processing device may include a processing unit 1001 (e.g., a central processing unit, a graphics processing unit, etc.), which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 1002 or a program loaded from a storage device 1003 into a random access memory (RAM) 1004. The RAM 1004 also stores various programs and data required for the operation of the data processing device. The processing unit 1001, ROM 1002, and RAM 1004 are interconnected via a bus 1005. An input / output (I / O) interface 1006 is also connected to the bus. Typically, the following systems can be connected to I / O interface 1006: input devices 1007 including, for example, touchscreens, touchpads, keyboards, mice, image sensors, microphones, accelerometers, gyroscopes, etc.; output devices 1008 including, for example, liquid crystal displays (LCDs), speakers, vibrators, etc.; storage devices 1003 including, for example, magnetic tapes, hard disks, etc.; and communication devices 1009. Communication device 1009 allows the data processing device to communicate wirelessly or wiredly with other devices to exchange data. Although a data processing device with various systems is shown in the figure, it should be understood that it is not required to implement or possess all the systems shown. More or fewer systems may be implemented alternatively.
[0130] Specifically, according to the embodiments disclosed in this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments disclosed in this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via a communication device, or installed from storage device 1003, or installed from ROM 1002. When the computer program is executed by processing device 1001, it performs the functions defined in the methods of the embodiments disclosed in this application.
[0131] The data processing device provided in this application, employing the data processing method described in the above embodiments, can solve the technical problem of low efficiency in the existing manual processing of property insurance claims. Compared with the prior art, the beneficial effects of the data processing device provided in this application are the same as those of the data processing method provided in the above embodiments, and other technical features of the data processing device are the same as those disclosed in the previous embodiment method, and will not be repeated here.
[0132] It should be understood that the various parts disclosed in this application can be implemented using hardware, software, firmware, or a combination thereof. In the description of the above embodiments, specific features, structures, materials, or characteristics can be combined in any suitable manner in one or more embodiments or examples.
[0133] The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.
[0134] This application provides a computer-readable storage medium having computer-readable program instructions (i.e., a computer program) stored thereon, the computer-readable program instructions being used to execute the data processing method described in the above embodiments.
[0135] The computer-readable storage medium provided in this application may be, for example, a USB flash drive, but is not limited to, electrical, magnetic, optical, electromagnetic, infrared, or semiconductor systems, devices, or any combination thereof. More specific examples of computer-readable storage media may include, but are not limited to: electrical connections having one or more wires, portable computer disks, hard disks, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM or flash memory), optical fiber, portable compact disk read-only memory (CD-ROM), optical storage devices, magnetic storage devices, or any suitable combination thereof. In this embodiment, the computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, system, or device. The program code contained on the computer-readable storage medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (Radio Frequency), etc., or any suitable combination thereof.
[0136] The aforementioned computer-readable storage medium may be included in a data processing device or may exist independently without being assembled into a data processing device.
[0137] Computer program code for performing the operations of this application can be written in one or more programming languages or a combination thereof, including object-oriented programming languages such as Java, Smalltalk, and C++, and conventional procedural programming languages such as the "C" language or similar programming languages. The program code can be executed entirely on the user's computer, partially on the user's computer, as a standalone software package, partially on the user's computer and partially on a remote computer, or entirely on a remote computer or server. In cases involving remote computers, the remote computer can be connected to the user's computer via any type of network—including a Local Area Network (LAN) or a Wide Area Network (WAN)—or can be connected to an external computer (e.g., via the Internet using an Internet service provider).
[0138] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. In this regard, each block in a flowchart or block diagram may represent a module, segment, or portion of code containing one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in the block diagrams and / or flowcharts, and combinations of blocks in the block diagrams and / or flowcharts, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.
[0139] The modules described in the embodiments of this application can be implemented in software or hardware. The names of the modules do not necessarily limit the functionality of the unit itself.
[0140] The readable storage medium provided in this application is a computer-readable storage medium that stores computer-readable program instructions (i.e., a computer program) for executing the above-described data processing method, which can solve the technical problem of low efficiency in the existing manual processing of property insurance claims. Compared with the prior art, the beneficial effects of the computer-readable storage medium provided in this application are the same as the beneficial effects of the data processing method provided in the above embodiments, and will not be repeated here.
[0141] This application also provides a computer program product, including a computer program that, when executed by a processor, implements the steps of the data processing method described above.
[0142] The computer program product provided in this application can solve the technical problem of low efficiency in the existing manual processing of property insurance claims. Compared with the prior art, the beneficial effects of the computer program product provided in this application are the same as the beneficial effects of the data processing method provided in the above embodiments, and will not be repeated here.
[0143] The above description is only a part of the embodiments of this application and does not limit the patent scope of this application. All equivalent structural transformations made under the technical concept of this application and using the contents of the specification and drawings of this application, or direct / indirect applications in other related technical fields, are included in the patent protection scope of this application.
Claims
1. A data processing method, characterized in that, The data processing method includes the following steps: Extract key claim elements from the claim information submitted by the user, and determine the initial claim settlement plan based on the key claim elements; Obtain the accident environment information corresponding to the claim information, verify the key elements of the claim based on the accident environment information, and obtain the verification result; If the verification result is successful, the initial claim settlement plan is modified based on the accident environment information to obtain the target claim settlement plan.
2. The data processing method as described in claim 1, characterized in that, After the step of revising the initial claim settlement plan based on the accident environment information to obtain the target claim settlement plan when the verification result is successful, the method further includes: Determine the confidence level of the target claims settlement plan; If the confidence level meets the preset confidence level conditions, the accident information is verified through a smart contract, and compensation is paid to the user after the verification is successful.
3. The data processing method as described in claim 1, characterized in that, The steps of extracting key claim elements from user-submitted claim information and determining an initial claim settlement plan based on these key claim elements include: The user-submitted claim information is converted into text format to obtain the claim information. Entity recognition is performed based on the claim information in the text format to obtain the accident entity recognition result; The key elements of the claim are determined based on the accident entity identification results and the claim information in the text format. The initial claims settlement plan is determined based on the aforementioned key elements of claims settlement and the preset compensation model.
4. The data processing method as described in claim 3, characterized in that, Before the step of determining the initial claims settlement plan based on the key claims elements and the preset compensation model, the method further includes: Obtain policy recall information, and determine sample data and policy recall rate based on the policy recall information; The initial compensation model is trained based on the sample data and the policy recall rate to obtain the preset compensation model.
5. The data processing method according to any one of claims 1-4, characterized in that, The step of verifying the key elements of the claim based on the accident environment information and obtaining the verification result includes: Feature extraction is performed on the accident environment information to obtain accident environment features; Determine the accident environment feature vector based on the aforementioned accident environment characteristics; Determine the feature vector of the key claim elements based on the aforementioned key claim elements; Determine the degree of matching between the accident environment feature vector and the feature vector of the key elements of the claims settlement; The verification results of the key elements of the claim are determined based on the matching degree.
6. The data processing method according to any one of claims 1-4, characterized in that, The step of revising the initial claim settlement plan based on the accident environment information to obtain the target claim settlement plan when the verification result is successful includes: If the verification result is successful, the driver's degree of responsibility is determined based on the accident environment information; The initial claim settlement plan is modified based on the liability ratio to obtain the target claim settlement plan.
7. A data processing apparatus, characterized in that, The data processing device includes: The initial claims settlement plan determination module is used to extract key claims elements from the claims information submitted by the user and determine the initial claims settlement plan based on the key claims elements. The verification module is used to obtain the accident environment information corresponding to the claim information, verify the key elements of the claim based on the accident environment information, and obtain the verification result. The initial claims settlement plan modification module is used to modify the initial claims settlement plan based on the accident environment information when the verification result is successful, so as to obtain the target claims settlement plan.
8. A data processing device, characterized in that, The device includes: a memory, a processor, and a computer program stored in the memory and executable on the processor, the computer program being configured to implement the steps of the data processing method as described in any one of claims 1 to 6.
9. A storage medium, characterized in that, The storage medium is a computer-readable storage medium, and a computer program is stored on the storage medium. When the computer program is executed by a processor, it implements the steps of the data processing method as described in any one of claims 1 to 6.
10. A computer program product, characterized in that, The computer program product includes a computer program that, when executed by a processor, implements the steps of the data processing method as described in any one of claims 1 to 6.