Typhoon claim settlement disaster damage assessment method, device, equipment and medium

Abstract

The invention relates to the technical field of artificial intelligence, is applied to financial or medical insurance claim settlement scenes, and particularly relates to a typhoon claim settlement disaster damage assessment method, device and equipment and a medium, and the method comprises the steps: obtaining a multi-source typhoon data set, carrying out the feature data recognition of the multi-source typhoon data set, and constructing a target claim settlement knowledge base according to the feature data; and inputting the obtained real-time typhoon data and early warning city information into the target claim settlement evaluation model based on the target claim settlement knowledge base to obtain a claim settlement evaluation result. Visibly, according to the method, feature data identification is performed on the multi-source typhoon data set, the target claim knowledge base is constructed according to the feature data, and then the acquired real-time typhoon data and early warning city information are input into the target claim evaluation model according to the target claim knowledge base, so that accurate evaluation of typhoon claim disaster damage is realized; and the insurance claim settlement efficiency is improved, so that an insurance claim settlement reference for typhoon disasters is provided for an insurance company.

Description

Technical Field

[0001] This invention relates to the field of artificial intelligence technology, and in particular to a method, apparatus, equipment and medium for assessing typhoon disaster damage in claims processing. Background Technology

[0002] Typhoons, as a common natural disaster, often cause severe property damage and health injuries to coastal areas and even some inland regions, posing a significant challenge to corporate property insurance and medical insurance claims. For insurance companies, conducting damage assessments for impending typhoons is crucial for making advance preparations for disaster prevention and mitigation, and for conducting claims processing quickly and efficiently. Currently, most existing methods for assessing typhoon damage in claims rely on traditional statistical models. However, these models only consider a single data source and struggle to handle complex multi-source spatiotemporal data. They fail to fully account for the combined impact of factors such as typhoons, rainfall, and urban disaster preparedness, resulting in low accuracy and poor timeliness. Consequently, the assessment results are often biased and cannot provide valuable information for future claims processing. Therefore, improving the accuracy of typhoon damage assessment and providing a reliable reference for typhoon disaster claims is a pressing technical challenge. Summary of the Invention

[0003] Therefore, in response to the above-mentioned technical problems, this invention provides a method, apparatus, equipment, and medium for assessing typhoon disaster damage, which can improve the accuracy of typhoon disaster damage assessment and thus provide a reference for typhoon disaster claims.

[0004] The first aspect of this application provides a method for assessing typhoon damage in claims, the method comprising: Obtain a multi-source typhoon dataset, which includes historical typhoon path data, meteorological precipitation data, typhoon random simulation data, and claims data; Feature data identification is performed on the multi-source typhoon dataset, and a target claims knowledge base is constructed based on the feature data. The feature data includes the degree of typhoon impact, rainfall data, urban disaster prevention capabilities, and the amount of claims losses. Based on the target claims knowledge base, the acquired real-time typhoon data and warning city information are input into the target claims assessment model to obtain the claims assessment results. The typhoon claims assessment results include the assessment of the amount of claims loss, the reasons for the assessment, and the assessment confidence level.

[0005] A second aspect of this application provides a typhoon claims damage assessment device, the typhoon claims damage assessment device comprising: The acquisition module is used to acquire a multi-source typhoon dataset, wherein the multi-source typhoon dataset includes historical typhoon path data, meteorological precipitation data, typhoon random simulation data, and claims case data; The construction module is used to identify feature data in the multi-source typhoon dataset and construct a target claims knowledge base based on the feature data. The feature data includes the degree of typhoon impact, rainfall data, urban disaster prevention capabilities, and the amount of claims loss. The assessment module is used to input the acquired real-time typhoon data and warning city information into the target claims assessment model based on the target claims knowledge base to obtain the claims assessment results. The typhoon claims assessment results include the assessment of the amount of claims loss, the reasons for the assessment, and the assessment confidence level.

[0006] Thirdly, a computer device is provided, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor, when executing the computer program, implements the typhoon claims damage assessment method as described in the first aspect.

[0007] Fourthly, a computer-readable storage medium is provided, the computer-readable storage medium storing a computer program, which, when executed by a processor, implements the typhoon claims disaster loss assessment method as described in the first aspect.

[0008] In summary, this invention provides a method, apparatus, equipment, and medium for assessing typhoon claims damage. It acquires a multi-source typhoon dataset, including historical typhoon path data, meteorological precipitation data, typhoon random simulation data, and claims case data. Feature data is identified within the multi-source typhoon dataset, and a target claims knowledge base is constructed based on this feature data. The feature data includes the degree of typhoon impact, rainfall data, urban disaster prevention capabilities, and the amount of claims damage. Based on this target claims knowledge base, the acquired real-time typhoon data and early warning city information are input into the target claims assessment model to obtain the claims assessment result. The typhoon claims assessment result includes an assessment of the amount of claims damage, the reasons for the assessment, and the assessment confidence level. As can be seen, this application identifies the features of multi-source typhoon datasets and constructs a target claims knowledge base based on these features. Then, based on this target claims knowledge base, it inputs the acquired real-time typhoon data and early warning city information into the target claims assessment model, thereby achieving accurate assessment of typhoon claims damage and improving insurance claims efficiency, thus providing insurance companies with a reference for typhoon disaster insurance claims. Attached Figure Description

[0009] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments of the present invention will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0010] Figure 1 This is a schematic diagram of an application environment for a typhoon claims damage assessment method provided in an embodiment of the present invention; Figure 2 This is a flowchart illustrating a typhoon damage assessment method according to an embodiment of the present invention. Figure 3 This is a schematic diagram of the structure of a typhoon claims damage assessment device provided in an embodiment of the present invention; Figure 4 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. Detailed Implementation

[0011] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0012] It should be understood that, when used in this specification and the appended claims, the term "comprising" indicates the presence of the described features, integrals, steps, operations, elements and / or components, but does not exclude the presence or addition of one or more other features, integrals, steps, operations, elements, components and / or collections thereof.

[0013] It should also be understood that the term “and / or” as used in this specification and the appended claims refers to any combination of one or more of the associated listed items and all possible combinations, and includes such combinations.

[0014] As used in this specification and the appended claims, the term "if" may be interpreted, depending on the context, as "when," "once," or "in response to determination." Similarly, the phrase "if determined" or "if matched to [described condition or event]" may be interpreted, depending on the context, as meaning "once determined," "in response to determination," "once matched to [described condition or event]," or "in response to matched to [described condition or event]."

[0015] Furthermore, in the description of this invention and the appended claims, the terms "first," "second," "third," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0016] References to "one embodiment" or "some embodiments" as described in this specification mean that one or more embodiments of the invention include a specific feature, structure, or characteristic described in connection with that embodiment. Therefore, the phrases "in one embodiment," "in some embodiments," "in other embodiments," "in still other embodiments," etc., appearing in different parts of this specification do not necessarily refer to the same embodiment, but rather mean "one or more, but not all, embodiments," unless otherwise specifically emphasized. The terms "comprising," "including," "having," and variations thereof mean "including but not limited to," unless otherwise specifically emphasized.

[0017] It should be understood that the sequence number of each step in the following embodiments does not imply the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present invention.

[0018] To illustrate the technical solution of the present invention, specific embodiments are described below.

[0019] An embodiment of the present invention provides a method for assessing typhoon damage in claims processing, which can be applied to, for example... Figure 1 In this application environment, the terminal device communicates with the server. The terminal device includes, but is not limited to, PDAs, desktop computers, laptops, ultra-mobile personal computers (UMPCs), netbooks, and personal digital assistants (PDAs). The server can be implemented using a standalone server or a server cluster consisting of multiple servers.

[0020] This application can be used in the fields of financial insurance claims and medical insurance claims. In the financial insurance claims field, taking the severe damage to a photovoltaic power station of an energy group caused by Typhoon Mangkhut in Wenchang, Hainan as an example, the insurance company, based on the assessment results, provided property insurance compensation for the photovoltaic power station during the typhoon, effectively reducing the potential risk of loss. In the medical insurance claims field, taking the dengue fever outbreak in parts of Hainan due to flooding after Typhoon Mangkhut as an example, the insurance company, based on the assessment results, provided health insurance compensation to patients, ensuring timely payment of medical expenses.

[0021] See Figure 2This is a flowchart illustrating a typhoon damage assessment method according to an embodiment of the present invention. Figure 2 As shown, this typhoon damage assessment method can be implemented through the following steps.

[0022] S201: Obtain a multi-source typhoon dataset, wherein the multi-source typhoon dataset includes historical typhoon path data, meteorological precipitation data, typhoon random simulation data, and claims case data.

[0023] In step S201, historical typhoon track data records the spatial location and time series data of past typhoons throughout their entire life cycle, from formation and development to landfall and dissipation. This data forms the basis for analyzing typhoon movement patterns and impact ranges; here, historical typhoon track data from 1949 to 2024 is used as an example. Meteorological precipitation data consists of real-time observation and forecast data related to rainfall accompanying typhoons, covering the intensity, distribution, and duration of rainfall in the typhoon-affected area. This data is crucial for assessing secondary disasters caused by heavy rain; here, precipitation data from national meteorological observation stations since 2000 is used as an example. Typhoon stochastic simulation data is virtual typhoon event data generated through meteorological models and statistical algorithms. This data compensates for the insufficient sample size of historical typhoon data and expands the coverage of risk assessment. Here, a 10,000-year stochastic simulation event set of typhoons in the western Pacific Ocean, generated based on the STORM typhoon simulation method, is used. Claims data comprises past typhoon-related corporate property / personal insurance claims records processed by insurance companies; this data is key business data connecting meteorological data and actual losses; here, corporate property insurance claims data since 2000 is used as an example.

[0024] In this embodiment, historical typhoon track data can be sourced from the China Meteorological Administration Tropical Cyclone Data Center (CMA), and the specific time period can be set according to actual conditions; this application does not impose any limitations on this. Meteorological precipitation data can be sourced from meteorological satellite remote sensing, and claims data can be sourced from insurance company internal databases, industry claims reports, etc. It is understandable that due to the limited number of historical typhoon track data due to time constraints, the subsequent claims damage assessment results are not comprehensive enough. Therefore, simulated random typhoon data can also be obtained to expand the typhoon dataset. Specifically, the random typhoon data is typhoon random simulation data generated based on the STORM (Synthetic Tropical Cyclic One Gene Ration Model) model. It should be noted that the STORM model uses statistical methods to randomly simulate typhoon events using climatic forcing fields, typhoon track, intensity, and formation location data. By integrating the obtained historical typhoon track data, meteorological precipitation data, typhoon random simulation data, and claims data, a multi-source typhoon dataset is obtained. By taking the steps described above, the limitations of single-source data are overcome, and a more comprehensive multi-source typhoon dataset is integrated, laying the foundation for improving the accuracy of typhoon damage assessment in subsequent claims.

[0025] S102: Perform feature data identification on the multi-source typhoon dataset and construct a target claims knowledge base based on the feature data, wherein the feature data includes the degree of typhoon impact, rainfall data, urban disaster prevention capabilities, and the amount of claims losses.

[0026] In step S102, the typhoon impact level measures the typhoon's destructive potential to the affected area and is the primary driving factor for disaster losses. Rainfall data refers to the intensity, duration, and coverage of precipitation accompanying the typhoon, which is a key trigger for secondary disasters. Urban disaster prevention capacity refers to the hardware (flood control dikes, drainage systems) and software (emergency response) levels of the affected city in resisting typhoon disasters. The amount of compensation loss is the amount actually paid by the insurance institution after the typhoon disaster. By identifying the feature data of the multi-source typhoon dataset, that is, extracting key information dimensions that have a direct or indirect impact on the compensation losses caused by the typhoon from the multi-source typhoon dataset (historical typhoon path data, meteorological precipitation data, typhoon random simulation data, and claims case data), and transforming them into standardized, computable feature data, statistical analysis (such as correlation analysis, regression analysis) or machine learning (such as decision trees, XGBoost) is used to explore the causal relationships between features. An appropriate storage architecture is selected, and a target claims knowledge base is constructed based on the causal relationships between features.

[0027] In one embodiment of the invention, feature data identification is performed on the multi-source typhoon dataset, and a target claims knowledge base is constructed based on the feature data, including: The historical typhoon path data is analyzed to determine the typhoon intensity, typhoon landfall information, and spatial correlation of typhoon path cities in the target affected cities. The target affected cities are all affected cities within the geographical area covered by extending 5 degrees of latitude or longitude to both sides of the typhoon center's movement path as the baseline. The degree of typhoon impact in the target affected cities is determined based on the typhoon intensity, typhoon landfall information, and spatial correlation of cities along the typhoon path in the target affected cities. The meteorological precipitation data is analyzed to determine the rainfall data in the target affected city. The rainfall data includes the maximum single-day rainfall during the typhoon, the maximum cumulative rainfall over a continuous period during the typhoon, and the maximum number of consecutive days of heavy rain during the typhoon. Based on the aforementioned typhoon random simulation data, long-term typhoon characteristics were calculated. Based on the pre-defined first mapping relationship and the long-term typhoon characteristics, the urban disaster prevention capabilities in the target affected cities are determined. The first mapping relationship is the correspondence between the pre-defined long-term typhoon characteristics and the urban disaster prevention capabilities. The urban disaster prevention capabilities include the distribution of simulated typhoon levels experienced by the city, the simulated typhoon landfall experienced by the city, and the simulated landfall level distribution experienced by the city. The claims data and historical typhoon events are correlated in multiple dimensions, and the correlated claims data are aggregated and calculated according to the typhoon event type to obtain the amount of claims loss in the target affected city. Based on the typhoon's impact, rainfall data, urban disaster prevention capabilities, and the amount of claims losses, an initial claims knowledge base is constructed. The initial claims knowledge base is validated based on preset validation rules to obtain the target claims knowledge base.

[0028] Specifically, by analyzing historical typhoon path data, the spatial correlation of typhoon intensity, landfall information, and typhoon path cities in the target affected cities is determined. The target affected cities are all cities within the geographical area covered by extending 5 degrees of latitude / longitude to both sides of the typhoon's center path as the baseline. For example, if the typhoon path moves from east to west along 25°N, the target affected cities are all cities within the range of 20°N-30°N. Typhoon intensity is defined as the maximum level and maximum wind speed of the typhoon along its path, and the level and wind speed of the typhoon closest to the affected city. Typhoon landfall information includes the city where the typhoon makes landfall, the typhoon's landfall level, and wind speed. Spatial correlation of cities along the typhoon's path includes whether the typhoon's path passes through the city, the maximum level and wind speed when the typhoon passes through, and the percentage of the city's area covered by the typhoon's path (between 0 and 1). Based on the typhoon intensity, landfall information, and spatial correlation of cities along the typhoon's path in the target affected city, the degree of typhoon impact in the target affected city is determined. By analyzing meteorological precipitation data, rainfall data in the target affected city is determined. This rainfall data includes the maximum single-day rainfall, the maximum cumulative rainfall over a continuous period during the typhoon period, and the maximum number of consecutive days of heavy rain during the typhoon period. The maximum single-day rainfall is the maximum value of the daily cumulative rainfall during the typhoon period; the maximum cumulative rainfall over a continuous period is the maximum value of the cumulative rainfall over 12 / 24 / 36 hours calculated using a sliding calculation; and the maximum number of consecutive days of heavy rain is the number of consecutive days with rainfall ≥50mm.

[0029] Then, by analyzing the random typhoon simulation data, long-term typhoon characteristics are calculated. Specifically, based on a 10,000-year typhoon simulation dataset, long-term typhoon characteristics are calculated. It can be considered that the more frequently a city experiences typhoons of a certain level, the stronger its disaster prevention capability and the lower the damage level for that level of typhoon. A pre-defined correspondence between long-term typhoon characteristics and urban disaster prevention capabilities is established. Based on this pre-defined first mapping relationship and the long-term typhoon characteristics, the urban disaster prevention capability in the target affected city is determined. Urban disaster prevention capability includes the distribution of simulated typhoon levels experienced by the city, the total number of typhoons and their annual average frequency corresponding to each typhoon level, and the landfall and landfall level of simulated typhoons experienced by the city. The system performs multi-dimensional correlation analysis between claims data and corresponding historical typhoon events. It then aggregates and calculates the associated claims data according to typhoon event type to determine the amount of loss in the target affected cities. The multi-dimensional correlation includes temporal and spatial correlation. Temporal correlation, considering the delay in reporting, is defined as the incident occurring between the typhoon's start time and 10 days after its end, with the cause of the incident listed as related to the typhoon or heavy rain. Spatial correlation is determined by parsing the incident address into a standardized province-city-district / county format. If the city where the incident address is located is within the list of cities affected by the typhoon's path, then the case is considered to have spatial correlation. Then, based on the typhoon's impact, rainfall data, urban disaster prevention capabilities, and the amount of claims losses, an initial claims knowledge base is constructed. This initial knowledge base is then validated using pre-defined validation rules, including data integrity rules, logical consistency rules, and data accuracy rules. Specifically, the data integrity rule requires that each knowledge entry contain all core fields; if one or more fields are missing, it is marked as "to be supplemented." The logical consistency rule requires consistency between typhoon intensity and claims losses (e.g., if a typhoon has a wind force of level 8 but claims losses far exceed the historical average for the same intensity, verification is required); consistency between rainfall data and loss types (e.g., if the maximum single-day rainfall is 50mm but "waterlogging losses" account for over 80%, the labeling of the loss cause needs to be investigated for errors); and the data accuracy rule requires that the typhoon intensity and rainfall data in each knowledge entry be compared with authoritative meteorological data sources (e.g., the National Meteorological Information Center), and corrections are made if the error exceeds 10%. The validated initial claims knowledge base is then used as the target claims knowledge base. It is evident that by covering the entire chain of typhoon, rainfall, disaster prevention, and loss, a target claims knowledge base can be effectively constructed. This enables accurate typhoon claims damage assessment, reduces errors, significantly improves the efficiency of claims processing for users, and ensures the timeliness and flexibility of claims business applications.

[0030] In this embodiment, by identifying the feature data of multi-source typhoon datasets, a target claims knowledge base can be constructed efficiently and accurately based on the feature data, eliminating data silos and realizing the value transformation of multi-source data into unified knowledge. This provides strong support for improving the accuracy and efficiency of typhoon claims damage assessment, improves claims settlement efficiency, and reduces reliance on manual labor and assessment errors.

[0031] S203: Based on the target claims knowledge base, the acquired real-time typhoon data and warning city information are input into the target claims assessment model to obtain the claims assessment results, wherein the typhoon claims assessment results include the assessment of the amount of claims loss, the reasons for the assessment, and the assessment confidence level.

[0032] In step S203, by analyzing historical typhoon path data, the typhoon intensity, typhoon landfall information, and spatial correlation of cities along the typhoon path in the target affected cities are determined. The target affected cities are all cities within the geographical area covered by extending 5 degrees of latitude / longitude to both sides of the typhoon's center path as the baseline. For example, if the typhoon path moves from east to west along 25°N, the target affected cities are all cities within the range of 20°N-30°N. Typhoon intensity is defined as the maximum level and maximum wind speed of the typhoon along its path, and the level and wind speed of the typhoon closest to the affected city. Typhoon landfall information includes the city where the typhoon makes landfall, the typhoon's landfall level, and wind speed. Spatial correlation of cities along the typhoon's path includes whether the typhoon's path passes through the city, the maximum level and wind speed when the typhoon passes through, and the percentage of the city's area covered by the typhoon's path (between 0 and 1). Based on the typhoon intensity, landfall information, and spatial correlation of cities along the typhoon's path in the target affected city, the degree of typhoon impact in the target affected city is determined. By analyzing meteorological precipitation data, rainfall data in the target affected city is determined. This rainfall data includes the maximum single-day rainfall, the maximum cumulative rainfall over a continuous period during the typhoon period, and the maximum number of consecutive days of heavy rain during the typhoon period. The maximum single-day rainfall is the maximum value of the daily cumulative rainfall during the typhoon period; the maximum cumulative rainfall over a continuous period is the maximum value of the cumulative rainfall over 12 / 24 / 36 hours calculated using a sliding calculation; and the maximum number of consecutive days of heavy rain is the number of consecutive days with rainfall ≥50mm.

[0033] Then, by analyzing the random typhoon simulation data, long-term typhoon characteristics are calculated. Specifically, based on a 10,000-year typhoon simulation dataset, long-term typhoon characteristics are calculated. It can be considered that the more frequently a city experiences typhoons of a certain level, the stronger its disaster prevention capability and the lower the damage level for that level of typhoon. A pre-defined correspondence between long-term typhoon characteristics and urban disaster prevention capabilities is established. Based on this pre-defined first mapping relationship and the long-term typhoon characteristics, the urban disaster prevention capability in the target affected city is determined. Urban disaster prevention capability includes the distribution of simulated typhoon levels experienced by the city, the total number of typhoons and their annual average frequency corresponding to each typhoon level, and the landfall and landfall level of simulated typhoons experienced by the city. The system performs multi-dimensional correlation analysis between claims data and corresponding historical typhoon events. It then aggregates and calculates the associated claims data according to typhoon event type to determine the amount of loss in the target affected cities. The multi-dimensional correlation includes temporal and spatial correlation. Temporal correlation, considering the delay in reporting, is defined as the incident occurring between the typhoon's start time and 10 days after its end, with the cause of the incident listed as related to the typhoon or heavy rain. Spatial correlation is determined by parsing the incident address into a standardized province-city-district / county format. If the city where the incident address is located is within the list of cities affected by the typhoon's path, then the case is considered to have spatial correlation. Then, based on the typhoon's impact, rainfall data, urban disaster prevention capabilities, and the amount of claims losses, an initial claims knowledge base is constructed. This initial knowledge base is then validated using pre-defined validation rules, including data integrity rules, logical consistency rules, and data accuracy rules. Specifically, the data integrity rule requires that each knowledge entry contain all core fields; if one or more fields are missing, it is marked as "to be supplemented." The logical consistency rule requires consistency between typhoon intensity and claims losses (e.g., if a typhoon has a wind force of level 8 but claims losses far exceed the historical average for the same intensity, verification is required); consistency between rainfall data and loss types (e.g., if the maximum single-day rainfall is 50mm but "waterlogging losses" account for over 80%, the labeling of the loss cause needs to be investigated for errors); and the data accuracy rule requires that the typhoon intensity and rainfall data in each knowledge entry be compared with authoritative meteorological data sources (e.g., the National Meteorological Information Center), and corrections are made if the error exceeds 10%. The validated initial claims knowledge base is then used as the target claims knowledge base. It is evident that by covering the entire chain of typhoon, rainfall, disaster prevention, and loss, a target claims knowledge base can be effectively constructed. This enables accurate typhoon claims damage assessment, reduces errors, significantly improves the efficiency of claims processing for users, and ensures the timeliness and flexibility of claims business applications.

[0034] In one embodiment of the invention, the target claims assessment model is established as follows: Multiple claim types can be pre-defined; Based on a pre-built claims database, acquire claims data related to any one of the multiple claims types, and use the claims data as a training dataset; Based on the claims data, obtain the actual claims data corresponding to the claims type, and use the actual claims data as the validation set corresponding to the training dataset; The training dataset is input into the initial claims assessment model to perform a forward operation, thereby obtaining the initial claims assessment result corresponding to the claims type; The initial claims assessment results are fitted to the validation set to obtain the degree of fit; Determine whether the fit degree is less than a preset fit threshold; If the fitting degree is less than the preset fitting threshold, the initial claims assessment model is iteratively trained according to the loss function of the initial claims assessment model until the trained initial claims assessment model meets the expected performance, and the target claims assessment model is obtained.

[0035] Specifically, multiple claim types are pre-defined, categorized as: political, typhoon disaster, public health, and economic. Sudden outbreaks of these types can drastically alter daily claims processing. For example, claims processing volume increases significantly during a typhoon disaster. Taking the typhoon disaster type as an example, multiple relevant claim data points are retrieved from the claims database and used as training samples. In machine learning and pattern recognition, samples are typically divided into three independent parts: a training set, a validation set, and a test set. The training set is used to estimate the model, the validation set is used to determine network structure or parameters controlling model complexity, and the test set is used to evaluate the performance of the ultimately selected optimal model. This application trains an initial model based on a training dataset and a validation set. It obtains actual claim data corresponding to different claim types from claim data and uses this actual claim data as the validation set corresponding to the training dataset. The training dataset is then input into the initial claim assessment model to perform forward computation, obtaining the initial claim assessment result corresponding to the claim type. The initial claim assessment result is then fitted to the validation set to obtain the goodness of fit. If the goodness of fit is less than a preset fitting threshold, the initial claim assessment model is iteratively trained based on the loss function in the initial claim assessment model until the trained initial claim assessment model meets the expected performance, thus obtaining the target claim assessment model. Fitting is the process of connecting a series of points on a plane with a smooth curve. Because this curve has countless possibilities, there are various fitting methods. The fitted curve can generally be represented by a function, and different functions have different fitting names. Common fitting methods include least squares curve fitting, and in MATLAB, polyfit can be used to fit polynomials. The goodness of fit refers to how well the regression line fits the observed values. The statistic for measuring goodness of fit is the coefficient of determination (R²). The maximum value of R² is 1. The closer the R² value is to 1, the better the regression line fits the observed values; conversely, the smaller the R² value, the worse the fit. R² measures the overall goodness of fit of the regression equation, expressing the overall relationship between the dependent variable and all independent variables. It should be noted that the preset fit threshold can be limited according to actual circumstances; this application does not impose any limitations on it. Through the above steps, noise interference is reduced, allowing the target claims assessment model to focus more on learning the patterns of the target type, thereby improving the accuracy of typhoon claims damage assessment. This enables advance disaster prevention and mitigation efforts and rapid claims processing for personnel in severely affected areas, improving claims efficiency.

[0036] In one embodiment of the invention, based on the target claims knowledge base, the acquired real-time typhoon data and warning city information are input into the target claims assessment model to obtain claims assessment results, including: Obtain the target city with the highest comprehensive disaster index within the preset typhoon time period; By conducting an online search based on typhoon tags, the real-time typhoon data and early warning city information corresponding to the target city for assessment are determined. Based on the path similarity between the real-time typhoon path and the historical typhoon path data in the real-time typhoon data, the top N historical similar typhoon samples with the highest similarity are selected, and the corresponding claim loss amount is retrieved in the target claims knowledge base, where N is a positive integer; Based on the real-time typhoon data and the early warning city information, indicator factor data is calculated, wherein the indicator factor data includes typhoon indicator factor data and rainstorm indicator factor data. Text processing is performed on the aforementioned indicator factor data to obtain typhoon text information and rainstorm text information; Based on the warning city information corresponding to the target city, the typhoon impact level and urban disaster prevention capabilities of the warning city information are retrieved from the target claims knowledge base. The data is combined with the target city, the real-time typhoon data, the early warning city information, the amount of compensation loss, the typhoon text information, the rainstorm text information, the degree of typhoon impact, and the city's disaster prevention capabilities. The combined data is then input into the target compensation assessment model to obtain the compensation assessment result.

[0037] Specifically, the assessment targets cities with the highest comprehensive disaster index within a preset typhoon timeframe are selected from all cities potentially affected by typhoons. This comprehensive disaster index is a quantitative indicator measuring the potential impact of a typhoon on a city, avoiding resource waste caused by indiscriminate analysis and focusing on high-risk areas. For example, if Typhoon "XX" is expected to affect 10 cities along the southeast coast, calculating the comprehensive disaster index reveals that Xiamen (index 85), Quanzhou (index 82), and Ningbo (index 78) rank in the top three. Xiamen is ultimately selected as the target city with the highest comprehensive disaster index. A network search is conducted using typhoon tags—that is, by typhoon name + international code—to determine the real-time typhoon data and warning information for the target city, ensuring the timeliness and authority of the assessment data and avoiding bias caused by outdated data. The path similarity between real-time typhoon paths and historical typhoon paths is calculated using the DTW algorithm or Hausdorff distance algorithm. This path similarity is quantified by spatial algorithms to determine the degree of overlap between real-time and historical typhoon paths (e.g., path overlap rate, consistency of movement direction, and overlap rate of affected areas). The paths are sorted in descending order of similarity, and the top N historical similar typhoon samples are selected. N is a positive integer representing the preset number of similar typhoon samples, usually N=3 or 5. Too few samples may lead to insufficient reference, while too many may introduce low-similarity interference samples. For example, if the real-time typhoon path has similarities of 92%, 88%, and 85% (N=3) to Typhoon Mangkhut (2018), Fireworks (2020), and Masang (2022), respectively, then these three historical typhoons are selected as samples.

[0038] Then, historical similar typhoon samples are input into the target claims knowledge base. The claims loss amount corresponding to the historical similar typhoon samples is retrieved from the target claims knowledge base. By calculating indicator factor data based on real-time typhoon data and warning city information, the indicator factor data includes typhoon indicator factor data and rainstorm indicator factor data. The indicator factor data is processed based on a preset structured text template to obtain typhoon text information and rainstorm text information. According to the warning city information corresponding to the target city, the typhoon impact degree and city disaster prevention capacity corresponding to the warning city information are retrieved from the target claims knowledge base. The target impact city, real-time typhoon data, warning city information, claims loss amount, typhoon text information, rainstorm text information, typhoon impact degree and city disaster prevention capacity are merged in sequence to obtain merged data. The merged data is then input into the target claims assessment model to obtain the claims assessment result. By integrating multi-source data, the assessment results become more comprehensive and the uncertainty is significantly reduced, thereby improving the accuracy of typhoon disaster loss assessment, reducing invalid analysis, and enabling the dispatch of personnel to severely affected areas in advance for disaster prevention and mitigation and rapid claims processing, thus improving claims efficiency.

[0039] In one embodiment of the invention, obtaining the target city with the highest comprehensive disaster index during a preset typhoon period includes: Obtain the list of typhoon warning cities that issued typhoon warning signals within the preset typhoon time period; For each typhoon-affected city, corresponding urban disaster assessment indicators are determined, a multi-level assessment indicator system is constructed, data for each assessment indicator is obtained, and the data is transformed into a form that can be used for analysis. The analytic hierarchy process (AHP) is used to calculate the weight coefficients of each indicator based on their relative importance. The comprehensive disaster index of each typhoon city is obtained by weighting and summing the indicator data with the corresponding weight coefficients. Based on the magnitude of the comprehensive disaster index of each typhoon city, the typhoon city with the highest comprehensive disaster index is selected as the target city for evaluation.

[0040] Specifically, by obtaining a list of typhoon-affected cities that issued typhoon warnings within a preset typhoon timeframe, corresponding urban disaster assessment indicators are determined for each city, constructing a multi-level assessment indicator system. Data for each assessment indicator is acquired and transformed into a form suitable for analysis. The urban disaster assessment indicators must cover the core dimensions of typhoon disaster. Then, using the analytic hierarchy process (AHP), the weight coefficients of each indicator are calculated based on their importance. For each typhoon-affected city, the indicator data is weighted and summed with the corresponding weight coefficients to obtain a comprehensive disaster index. The cities are then sorted in descending order based on their comprehensive disaster indices, and the city with the highest comprehensive disaster index is selected as the target city for assessment. Through these steps, only cities with actual risks are focused on, rather than indiscriminately assessing all cities. This reduces the cost of data collection and analysis, comprehensively reflects the true situation of urban disasters, and improves the efficiency of urban disaster assessment, thus providing decision support for post-disaster relief and reconstruction.

[0041] In one embodiment of the invention, based on the path similarity between the real-time typhoon path and historical typhoon path data in the real-time typhoon data, the top N historical similar typhoon samples with the highest similarity are selected, including: Based on the real-time typhoon path and historical typhoon path data in the real-time typhoon data, multiple path similarities are calculated. Determine whether the similarity of each path satisfies a preset similarity threshold in turn; If the similarity of each path meets the preset similarity threshold, then the similarity of each path is sorted from high to low, and the top N historical similar typhoon samples are selected.

[0042] Specifically, the real-time typhoon paths and historical typhoon paths in the real-time typhoon data are aligned in both time and space dimensions. Based on the "temporal + spatial" characteristics of typhoon paths, a suitable spatial similarity algorithm is selected. The path similarity between the real-time typhoon paths and historical typhoon paths in the real-time typhoon data is calculated by using the DTW algorithm or the Hausdorff distance algorithm. Taking the DTW algorithm as an example, Dynamic Time Warping (DTW) is an algorithm used to measure the similarity between two time series (such as the latitude and longitude of typhoon path points changing over time). By stretching or compressing the time axis, the optimal matching path between the two series is found, and the minimum cumulative distance is calculated (the smaller the distance, the higher the similarity). The "minimum similarity threshold" is set based on business needs (such as claims assessment accuracy requirements and historical sample validity). For example, if the preset similarity threshold is 70%, and the real-time typhoon has a similarity of 85% with historical typhoon A (meets the threshold) but a similarity of 65% with historical typhoon B (does not meet the threshold), then only typhoon A is retained, and typhoon B is removed. This process is repeated for each path similarity, and if each path similarity meets the threshold, the paths with the highest similarity are sorted from highest to lowest. The top N historical typhoons are selected as the final historical similar typhoon sample. For example, if there are 8 similarities that meet the threshold (85%, 82%, 79%, 76%, 75%, 73%, 71%, 70%), and N=3, then the top 3 (85%, 82%, 79%) historical typhoons are selected as the historical similar typhoon sample. By following the above methods, we can avoid interference from low-similarity samples in subsequent analysis, ensure that the selected historical similar typhoon samples are closest to the real-time typhoon paths, lay the foundation for the accuracy and efficiency of typhoon claims damage assessment, and thus improve the claims settlement effect.

[0043] In one embodiment of the invention, after obtaining the claims assessment result, the process includes: The claims assessment results are input into the claims trend assessment model to determine the claims development trend of the current claims type. Determine whether the claim development trend of the claim type corresponding to the current claim is abnormal; If the claim development trend of the claim type corresponding to the current claim is abnormal, then determine the abnormal cause of the abnormal claim development trend of the claim type corresponding to the current claim; The cause of the anomaly is sent to the network device of the insurance monitoring center corresponding to the claim type, and the claim policy for the claim type is adjusted based on the cause of the anomaly. If the claim development trend of the claim type corresponding to the current claim is normal, the target risk level of the claim assessment result is evaluated by a preset claim risk model; The target claim recommendation corresponding to the target risk level is found in the pre-defined second mapping relationship, and the target claim recommendation is displayed. The second mapping relationship is a pre-configured correspondence between risk levels and claim recommendations.

[0044] Specifically, by inputting the claims assessment results into the claims trend assessment model, the claims development trend of the current claims type can be determined. This claims trend assessment model can be a time series analysis model (such as ARIMA or LSTM, used to analyze the periodicity and seasonality of historical trends) + a "multi-factor regression model" (incorporating external factors such as weather and policies to optimize trend judgment). Its training data consists of the past 3-5 years of categorized claims data + concurrent external data. The goal is to enable the claims trend assessment model to accurately identify normal and abnormal trends. Based on preset anomaly judgment criteria, the system verifies whether the claims development trend deviates from the normal range. That is, it judges whether the claims development trend corresponding to the current claims type is abnormal. The anomaly judgment criteria include: compared with the same period in history: for example, the estimated year-on-year increase of 40% in the amount of outpatient medical insurance claims in September far exceeds the year-on-year increase range of 5%-10% in the past 3 years, which is judged as abnormal; compared with the industry benchmark: for example, the estimated month-on-month increase of 20% in the claims volume of a certain region, while the average month-on-month increase of the claims volume of the whole province in the same period is only 5%, which is judged as abnormal; compared with the threshold: for example, the estimated rejection rate of critical illness insurance one-time claims reaches 15%, which exceeds the preset 10% abnormal threshold, and is judged as abnormal. If the current claim development trend for the corresponding claim type is abnormal, then any one claim data point is extracted from each claim data set in each claim dataset. Using the value of that claim data point in each claim data set as the observation value and the acquisition time of each claim data set as the horizontal axis, a time series corresponding to that claim data point is created. This results in multiple time series corresponding to multiple claim data points within that claim dataset. These multiple time series are processed to identify time series with inconsistent trends. The corresponding claim data point is then identified, and its corresponding claim indicator is marked as the cause of the abnormal claim development trend. This cause of the abnormality is sent to the network equipment of the monitoring center corresponding to the claim type, allowing for adjustments to the claim policy for that claim type. When an abnormal claim development trend occurs in the claim field, the cause of the abnormal trend is determined, and this cause is sent to the network-side equipment of the insurance monitoring center. This enables the insurance monitoring center to adjust the claim system accordingly based on the cause of the abnormality, optimize the claim system, improve the ability of the regulatory department to respond to abnormal claim areas, and enhance user claim satisfaction.If the current claim trend for the corresponding claim type is normal, a pre-set claim risk model is used to assess the target risk level of the claim assessment results. This pre-set claim risk model can be a classification model (such as Random Forest or XGBoost), and the training data consists of feature data of historical claim cases (amount, customer information, and material details) plus manually labeled risk levels. The goal is for the pre-set claim risk model to output an accurate risk level based on the input assessment results. Then, based on the second mapping relationship between pre-labeled risk levels and claim recommendations, the target claim recommendation corresponding to the target risk level is found and displayed. That is, the claim recommendation for high-risk and medium-risk levels can be set as not recommending claims, and the claim recommendation for low-risk levels can be set as recommending claims. For example, the target claim recommendation can be displayed through the display interface of a medical risk identification device. If the target claim recommendation is either not recommending claims or recommending claims, then the claims staff can process the user's claim application based on the target claim recommendation. It is evident that identifying potential risks through trend models, preventing the expansion of losses after the fact, optimizing claims policies in real time based on the causes of anomalies, and improving risk control flexibility can achieve dynamic management of claims risks, thereby improving claims efficiency.

[0045] In this embodiment, real-time typhoon data and early warning city information are input into the target claims assessment model based on the target claims knowledge base. This makes the claims assessment results more accurate and comprehensive, providing a more precise quantitative basis for typhoon claims damage assessment. This allows for the output of claims assessment and risk warnings as early as the disaster evolution stage, which is more timely than delayed statistics. By assessing the causes and contributions, it helps the claims, risk control, and operations teams quickly understand the sources of risk, thereby improving claims efficiency.

[0046] Taking the impact of Typhoon Yangliu on Raoping County, Guangdong Province, and the risk of flooding faced by aquaculture farmers as an example, this study obtains a multi-source typhoon dataset for the region and constructs a target claims knowledge base based on the characteristic data in the dataset. This allows the target claims assessment model to evaluate the claims outcome for Raoping County, Guangdong Province. Based on the assessed claims outcome, claims can then be processed through an insurance platform after Typhoon Yangliu has passed.

[0047] In summary, this invention provides a method, apparatus, equipment, and medium for assessing typhoon claims damage. It acquires a multi-source typhoon dataset, including historical typhoon path data, meteorological precipitation data, typhoon random simulation data, and claims case data. Feature data is identified within the multi-source typhoon dataset, and a target claims knowledge base is constructed based on this feature data. The feature data includes the degree of typhoon impact, rainfall data, urban disaster prevention capabilities, and the amount of claims damage. Based on this target claims knowledge base, the acquired real-time typhoon data and early warning city information are input into the target claims assessment model to obtain the claims assessment result. The typhoon claims assessment result includes an assessment of the amount of claims damage, the reasons for the assessment, and the assessment confidence level. As can be seen, this application identifies the features of multi-source typhoon datasets and constructs a target claims knowledge base based on these features. Then, based on this target claims knowledge base, it inputs the acquired real-time typhoon data and early warning city information into the target claims assessment model, thereby achieving accurate assessment of typhoon claims damage and improving insurance claims efficiency, thus providing insurance companies with a reference for typhoon disaster insurance claims.

[0048] Please see Figure 3 , Figure 3 This is a schematic diagram of the typhoon damage assessment device provided in an embodiment of the present invention. This typhoon damage assessment device corresponds one-to-one with the typhoon damage assessment method described in the above embodiments. Please refer to [link / reference] for details. Figure 2 as well as Figure 2 The relevant descriptions in the corresponding embodiments are shown below. For ease of explanation, only the parts relevant to this embodiment are shown. See also... Figure 3 The typhoon claims damage assessment device 30 includes: an acquisition module 31, a construction module 32, and an assessment module 33.

[0049] The acquisition module 31 is used to acquire a multi-source typhoon dataset, wherein the multi-source typhoon dataset includes historical typhoon path data, meteorological precipitation data, typhoon random simulation data, and claims case data; The construction module 32 is used to identify feature data in the multi-source typhoon dataset and construct a target claims knowledge base based on the feature data. The feature data includes the degree of typhoon impact, rainfall data, urban disaster prevention capabilities, and the amount of claims loss. The assessment module 33 is used to input the acquired real-time typhoon data and warning city information into the target claims assessment model based on the target claims knowledge base to obtain the claims assessment result, wherein the typhoon claims assessment result includes the assessment of the amount of claims loss, the assessment reason and the assessment confidence level.

[0050] Optionally, the above-mentioned construction module 32 is specifically used for: The historical typhoon path data is analyzed to determine the typhoon intensity, typhoon landfall information, and spatial correlation of typhoon path cities in the target affected cities. The target affected cities are all affected cities within the geographical area covered by extending 5 degrees of latitude or longitude to both sides of the typhoon center's movement path as the baseline. The degree of typhoon impact in the target affected cities is determined based on the typhoon intensity, typhoon landfall information, and spatial correlation of cities along the typhoon path in the target affected cities. The meteorological precipitation data is analyzed to determine the rainfall data in the target affected city. The rainfall data includes the maximum single-day rainfall during the typhoon, the maximum cumulative rainfall over a continuous period during the typhoon, and the maximum number of consecutive days of heavy rain during the typhoon. Based on the aforementioned typhoon random simulation data, long-term typhoon characteristics were calculated. Based on the pre-defined first mapping relationship and the long-term typhoon characteristics, the urban disaster prevention capabilities in the target affected cities are determined. The first mapping relationship is the correspondence between the pre-defined long-term typhoon characteristics and the urban disaster prevention capabilities. The urban disaster prevention capabilities include the distribution of simulated typhoon levels experienced by the city, the simulated typhoon landfall experienced by the city, and the simulated landfall level distribution experienced by the city. The claims data and historical typhoon events are correlated in multiple dimensions, and the correlated claims data are aggregated and calculated according to the typhoon event type to obtain the amount of claims loss in the target affected city. Based on the typhoon's impact, rainfall data, urban disaster prevention capabilities, and the amount of claims losses, an initial claims knowledge base is constructed. The initial claims knowledge base is validated based on preset validation rules to obtain the target claims knowledge base.

[0051] Optionally, the aforementioned evaluation module 33 is specifically used for: Multiple claim types can be pre-defined; Based on a pre-built claims database, acquire claims data related to any one of the multiple claims types, and use the claims data as a training dataset; Based on the claims data, obtain the actual claims data corresponding to the claims type, and use the actual claims data as the validation set corresponding to the training dataset; The training dataset is input into the initial claims assessment model to perform a forward operation, thereby obtaining the initial claims assessment result corresponding to the claims type; The initial claims assessment results are fitted to the validation set to obtain the degree of fit; Determine whether the fit degree is less than a preset fit threshold; If the fitting degree is less than the preset fitting threshold, the initial claims assessment model is iteratively trained according to the loss function of the initial claims assessment model until the trained initial claims assessment model meets the expected performance, and the target claims assessment model is obtained.

[0052] Optionally, the evaluation module 33 described above is also used for: Obtain the target city with the highest comprehensive disaster index within the preset typhoon time period; By conducting an online search based on typhoon tags, the real-time typhoon data and early warning city information corresponding to the target city for assessment are determined. Based on the path similarity between the real-time typhoon path and the historical typhoon path data in the real-time typhoon data, the top N historical similar typhoon samples with the highest similarity are selected, and the corresponding claim loss amount is retrieved in the target claims knowledge base, where N is a positive integer; Based on the real-time typhoon data and the early warning city information, indicator factor data is calculated, wherein the indicator factor data includes typhoon indicator factor data and rainstorm indicator factor data. Text processing is performed on the aforementioned indicator factor data to obtain typhoon text information and rainstorm text information; Based on the warning city information corresponding to the target city, the typhoon impact level and urban disaster prevention capabilities of the warning city information are retrieved from the target claims knowledge base. The data is combined with the target city, the real-time typhoon data, the early warning city information, the amount of compensation loss, the typhoon text information, the rainstorm text information, the degree of typhoon impact, and the city's disaster prevention capabilities. The combined data is then input into the target compensation assessment model to obtain the compensation assessment result.

[0053] Optionally, the evaluation module 33 described above is also used for: Obtain the list of typhoon warning cities that issued typhoon warning signals within the preset typhoon time period; For each typhoon-affected city, corresponding urban disaster assessment indicators are determined, a multi-level assessment indicator system is constructed, data for each assessment indicator is obtained, and the data is transformed into a form that can be used for analysis. The analytic hierarchy process (AHP) is used to calculate the weight coefficients of each indicator based on their relative importance. The comprehensive disaster index of each typhoon city is obtained by weighting and summing the indicator data with the corresponding weight coefficients. Based on the magnitude of the comprehensive disaster index of each typhoon city, the typhoon city with the highest comprehensive disaster index is selected as the target city for evaluation.

[0054] Optionally, the evaluation module 33 described above is also used for: Based on the real-time typhoon path and historical typhoon path data in the real-time typhoon data, multiple path similarities are calculated. Determine whether the similarity of each path satisfies a preset similarity threshold in turn; If the similarity of each path meets the preset similarity threshold, then the similarity of each path is sorted from high to low, and the top N historical similar typhoon samples are selected.

[0055] Optionally, the evaluation module 33 described above is then specifically used for: The claims assessment results are input into the claims trend assessment model to determine the claims development trend of the current claims type. Determine whether the claim development trend of the claim type corresponding to the current claim is abnormal; If the claim development trend of the claim type corresponding to the current claim is abnormal, then determine the abnormal cause of the abnormal claim development trend of the claim type corresponding to the current claim; The cause of the anomaly is sent to the network device of the insurance monitoring center corresponding to the claim type, and the claim policy for the claim type is adjusted based on the cause of the anomaly. If the claim development trend of the claim type corresponding to the current claim is normal, the target risk level of the claim assessment result is evaluated by a preset claim risk model; The target claim recommendation corresponding to the target risk level is found in the pre-defined second mapping relationship, and the target claim recommendation is displayed. The second mapping relationship is a pre-configured correspondence between risk levels and claim recommendations.

[0056] It should be noted that the information interaction and execution process between the above-mentioned units are based on the same concept as the method embodiments of the present invention. For details on their specific functions and technical effects, please refer to the method embodiments section, which will not be repeated here.

[0057] Figure 4 This is a schematic diagram of the structure of a computer device provided in an embodiment of the present invention. Figure 4 As shown, the computer device of this embodiment includes: at least one processor ( Figure 4 Only one is shown in the diagram), a memory, and a computer program stored in the memory and capable of running on at least one processor. When the processor executes the computer program, it implements the steps in any of the above embodiments of the typhoon compensation disaster damage assessment method.

[0058] This computer device may include, but is not limited to, a processor and memory. Those skilled in the art will understand that... Figure 4The examples of computer devices are merely examples and do not constitute a limitation on computer devices. Computer devices may include more or fewer components than shown in the illustration, or combinations of certain components, or different components, such as network interfaces, displays, and input systems.

[0059] In one embodiment, a computer-readable storage medium is provided, which, when executed by a processor in a computer device, enables the computer device to perform the steps of any embodiment of the typhoon compensation disaster damage assessment method disclosed in this invention, which will not be repeated here. The computer-readable storage medium may be non-volatile or volatile.

[0060] The processor referred to can be a CPU, but it can also be other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor can be a microprocessor or any conventional processor.

[0061] Memory includes readable storage media, internal memory, etc., wherein internal memory can be the RAM of a computer device, providing an environment for the operation of the operating system and computer-readable instructions stored in the readable storage media. The readable storage media can be the hard drive of the computer device, or in other embodiments, it can be an external storage device of the computer device, such as a plug-in hard drive, SmartMediaCard (SMC), SecureDigital (SD) card, or FlashCard. Furthermore, memory can include both internal storage units and external storage devices of the computer device. Memory is used to store the operating system, cooperative applications, bootloader, data, and other programs, such as program code of computer programs. Memory can also be used to temporarily store data that has been output or will be output.

[0062] Those skilled in the art will understand that implementing all or part of the processes in the above embodiments can be accomplished by a computer program instructing related hardware. This computer program can be stored in a non-volatile computer-readable storage medium. When executed, the computer program can include the processes of the embodiments described above. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in a variety of forms, such as static RAM (SRAM), dynamic RAM (DRAM), synchronous DRAM (SDRAM), dual data rate SDRAM (DDRSDRAM), enhanced SDRAM (ESDRAM), synchronous link DRAM (SLDRAM), RAMbus direct RAM (RDRAM), direct memory bus dynamic RAM (DRDRAM), and memory bus dynamic RAM (RDRAM).

[0063] Those familiar with the technical field will understand that, for ease of description and brevity, the above-described division of functional units and modules is used as an example. In practical applications, the above functions can be assigned to different functional units and modules as needed, that is, the internal structure of the system can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiments 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. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this invention. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here. 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.

[0064] The above-described embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should all be included within the protection scope of the present invention.

Claims

1. A method for assessing typhoon disaster damage in claims processing, characterized in that, include: Obtain a multi-source typhoon dataset, which includes historical typhoon path data, meteorological precipitation data, typhoon random simulation data, and claims data; Feature data identification is performed on the multi-source typhoon dataset, and a target claims knowledge base is constructed based on the feature data. The feature data includes the degree of typhoon impact, rainfall data, urban disaster prevention capabilities, and the amount of claims losses. Based on the target claims knowledge base, the acquired real-time typhoon data and warning city information are input into the target claims assessment model to obtain the claims assessment results. The typhoon claims assessment results include the assessment of the amount of claims loss, the reasons for the assessment, and the assessment confidence level.

2. The typhoon disaster loss assessment method as described in claim 1, characterized in that, The step of identifying feature data from the multi-source typhoon dataset and constructing a target claims knowledge base based on the feature data includes: The historical typhoon path data is analyzed to determine the typhoon intensity, typhoon landfall information, and spatial correlation of typhoon path cities in the target affected cities. The target affected cities are all affected cities within the geographical area covered by extending 5 degrees of latitude or longitude to both sides of the typhoon center's movement path as the baseline. The degree of typhoon impact in the target affected cities is determined based on the typhoon intensity, typhoon landfall information, and spatial correlation of cities along the typhoon path in the target affected cities. The meteorological precipitation data is analyzed to determine the rainfall data in the target affected city. The rainfall data includes the maximum single-day rainfall during the typhoon, the maximum cumulative rainfall over a continuous period during the typhoon, and the maximum number of consecutive days of heavy rain during the typhoon. Based on the aforementioned typhoon random simulation data, long-term typhoon characteristics were calculated. Based on the pre-defined first mapping relationship and the long-term typhoon characteristics, the urban disaster prevention capabilities in the target affected cities are determined. The first mapping relationship is the correspondence between the pre-defined long-term typhoon characteristics and the urban disaster prevention capabilities. The urban disaster prevention capabilities include the distribution of simulated typhoon levels experienced by the city, the simulated typhoon landfall experienced by the city, and the simulated landfall level distribution experienced by the city. The claims data and historical typhoon events are correlated in multiple dimensions, and the correlated claims data are aggregated and calculated according to the typhoon event type to obtain the amount of claims loss in the target affected city. Based on the typhoon's impact, rainfall data, urban disaster prevention capabilities, and the amount of claims losses, an initial claims knowledge base is constructed. The initial claims knowledge base is validated based on preset validation rules to obtain the target claims knowledge base.

3. The typhoon disaster loss assessment method as described in claim 1, characterized in that, The target claims assessment model is established in the following way: Multiple claim types can be pre-defined; Based on a pre-built claims database, acquire claims data related to any one of the multiple claims types, and use the claims data as a training dataset; Based on the claims data, obtain the actual claims data corresponding to the claims type, and use the actual claims data as the validation set corresponding to the training dataset; The training dataset is input into the initial claims assessment model to perform a forward operation, thereby obtaining the initial claims assessment result corresponding to the claims type; The initial claims assessment results are fitted to the validation set to obtain the degree of fit; Determine whether the fit degree is less than a preset fit threshold; If the fitting degree is less than the preset fitting threshold, the initial claims assessment model is iteratively trained according to the loss function of the initial claims assessment model until the trained initial claims assessment model meets the expected performance, and the target claims assessment model is obtained.

4. The typhoon disaster loss assessment method as described in claim 1, characterized in that, Based on the target claims knowledge base, the acquired real-time typhoon data and warning city information are input into the target claims assessment model to obtain claims assessment results, including: Obtain the target city with the highest comprehensive disaster index within the preset typhoon time period; By conducting an online search based on typhoon tags, the real-time typhoon data and early warning city information corresponding to the target city for assessment are determined. Based on the path similarity between the real-time typhoon path and the historical typhoon path data in the real-time typhoon data, the top N historical similar typhoon samples with the highest similarity are selected, and the corresponding claim loss amount is retrieved in the target claims knowledge base, where N is a positive integer; Based on the real-time typhoon data and the early warning city information, indicator factor data is calculated, wherein the indicator factor data includes typhoon indicator factor data and rainstorm indicator factor data. Text processing is performed on the aforementioned indicator factor data to obtain typhoon text information and rainstorm text information; Based on the warning city information corresponding to the target city, the typhoon impact level and urban disaster prevention capabilities of the warning city information are retrieved from the target claims knowledge base. The data is combined with the target city, the real-time typhoon data, the early warning city information, the amount of compensation loss, the typhoon text information, the rainstorm text information, the degree of typhoon impact, and the city's disaster prevention capabilities. The combined data is then input into the target compensation assessment model to obtain the compensation assessment result.

5. The typhoon disaster loss assessment method as described in claim 4, characterized in that, The method for identifying the target cities with the highest comprehensive disaster index within a preset typhoon time period includes: Obtain the list of typhoon warning cities that will issue typhoon warning signals within the preset typhoon time period; For each typhoon-affected city, corresponding urban disaster assessment indicators are determined, a multi-level assessment indicator system is constructed, data for each assessment indicator is obtained, and the data is transformed into a form that can be used for analysis. The analytic hierarchy process (AHP) is used to calculate the weight coefficients of each indicator based on their relative importance. The comprehensive disaster index of each typhoon city is obtained by weighting and summing the indicator data with the corresponding weight coefficients. Based on the magnitude of the comprehensive disaster index of each typhoon city, the typhoon city with the highest comprehensive disaster index is selected as the target city for evaluation.

6. The typhoon disaster loss assessment method as described in claim 4, characterized in that, The method of selecting the top N historical similar typhoon samples based on the path similarity between the real-time typhoon path and historical typhoon path data from the real-time typhoon data includes: Based on the real-time typhoon path and historical typhoon path data in the real-time typhoon data, multiple path similarities are calculated. Determine whether the similarity of each path satisfies a preset similarity threshold in turn; If the similarity of each path meets the preset similarity threshold, then the similarity of each path is sorted from high to low, and the top N historical similar typhoon samples are selected.

7. The typhoon disaster damage assessment method as described in claim 1, characterized in that, After obtaining the claims assessment results, the following are included: The claims assessment results are input into the claims trend assessment model to determine the claims development trend of the current claims type. Determine whether the claim development trend of the claim type corresponding to the current claim is abnormal; If the claim development trend of the claim type corresponding to the current claim is abnormal, then determine the abnormal cause of the abnormal claim development trend of the claim type corresponding to the current claim; The cause of the anomaly is sent to the network device of the insurance monitoring center corresponding to the claim type, and the claim policy for the claim type is adjusted based on the cause of the anomaly. If the claim development trend of the claim type corresponding to the current claim is normal, the target risk level of the claim assessment result is evaluated by a preset claim risk model; The target claim recommendation corresponding to the target risk level is found in the pre-defined second mapping relationship, and the target claim recommendation is displayed. The second mapping relationship is a pre-configured correspondence between risk levels and claim recommendations.

8. A typhoon damage assessment device for claims processing, characterized in that, include: The acquisition module is used to acquire a multi-source typhoon dataset, wherein the multi-source typhoon dataset includes historical typhoon path data, meteorological precipitation data, typhoon random simulation data, and claims case data; The construction module is used to identify feature data in the multi-source typhoon dataset and construct a target claims knowledge base based on the feature data. The feature data includes the degree of typhoon impact, rainfall data, urban disaster prevention capabilities, and the amount of claims loss. The assessment module is used to input the acquired real-time typhoon data and warning city information into the target claims assessment model based on the target claims knowledge base to obtain the claims assessment results. The typhoon claims assessment results include the assessment of the amount of claims loss, the reasons for the assessment, and the assessment confidence level.

9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the computer program, it implements the typhoon compensation disaster damage assessment method as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the typhoon compensation disaster damage assessment method as described in any one of claims 1 to 7.

Images