Method, device and equipment for processing abnormal insurance list and medium

Abstract

The invention belongs to the field of artificial intelligence and the field of financial science and technology, and relates to an insurance buying abnormity list processing method comprising the steps that an insurance buying abnormity list and corresponding historical insurance buying information are acquired, and the insurance buying abnormity list comprises a plurality of abnormal objects; and according to a preset screening rule, screening the insurance buying exception list to obtain a plurality of target exception objects. And classifying the plurality of target abnormal objects by using a pre-training classification model and a preset risk level strategy to determine a risk category. And then, according to a preset circulation rule and a risk category, determining a circulation path for each target abnormal object. And finally, transferring each target abnormal object and the historical insurance information thereof to a corresponding processing terminal according to a set path. The invention further provides a device, equipment and a medium. In addition, the invention also relates to a block chain technology, and the target historical insurance information can be stored in a block chain. The client resource utilization rate can be improved.

Description

Technical Field

[0001] The present application relates to the fields of artificial intelligence technology and financial technology, and in particular to a method, device, equipment and medium for processing an insurance exception list. Background Art

[0002] As the growth of my country's fuel vehicle market slows down, the auto insurance premium market is gradually becoming saturated, and property insurance companies are facing severe challenges in their auto-related premium business. In order to seek new growth points, non-auto property insurance has become an important breakthrough. However, in the actual promotion process, non-auto property insurance has encountered many difficulties.

[0003] On the one hand, customers have limited knowledge of non-auto insurance and are not clear about its coverage and importance, which leads to the fact that they often ignore its value when purchasing and prefer to choose familiar auto insurance products. This leads to a high initial service failure rate for non-auto property insurance and the loss of a large number of potential customers. On the other hand, traditional methods often simply give up on the list of initial service failures without in-depth mining and utilization. These lists contain rich potential customer information, but due to the lack of effective technical means to mine and analyze, a lot of resources are wasted. Summary of the invention

[0004] The purpose of the embodiments of the present application is to propose a method, device, equipment and medium for processing an insurance exception list, so as to solve the existing problem that after the initial service fails due to the insured's insufficient knowledge of non-auto insurance, the failure list resources are not effectively mined and utilized, resulting in low customer resource utilization.

[0005] In the first aspect, a method for processing an insurance abnormality list is provided, which adopts the following technical solution:

[0006] Obtain an insurance exception list and the target historical insurance information corresponding to the insurance exception list, wherein the insurance exception list includes multiple insurance exception objects; based on preset screening rules and target historical insurance information, screen the multiple abnormal insurance objects to obtain multiple screened target abnormal insurance objects; based on a preset risk level strategy and target historical insurance information, use a pre-trained classification model to classify the multiple target abnormal insurance objects and determine the risk category of each target abnormal insurance object; based on preset circulation rules and risk categories, determine the circulation path of each target abnormal insurance object; and circulate each target abnormal insurance object and target historical insurance information to the corresponding processing terminal according to the circulation path.

[0007] In the second aspect, a device for processing an insurance exception list is provided, which adopts the following technical solution:

[0008] The first acquisition module is used to acquire an insurance exception list and target historical insurance information corresponding to the insurance exception list, wherein the insurance exception list includes multiple insurance exception objects;

[0009] A screening module is used to screen multiple abnormal insurance objects based on preset screening rules and target historical insurance information to obtain multiple screened target abnormal insurance objects;

[0010] A classification module is used to classify multiple target abnormal insurance objects based on a preset risk level strategy and target historical insurance information, using a pre-trained classification model to determine the risk category of each target abnormal insurance object;

[0011] A path determination module is used to determine the flow path of each target abnormal insurance object based on preset flow rules and risk categories;

[0012] The circulation module is used to transfer each target abnormal insurance object and target historical insurance information to the corresponding processing terminal according to the circulation path.

[0013] In a third aspect, a computer device is provided, including a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, the steps of the method for processing the insurance exception list as described above are implemented.

[0014] In a fourth aspect, a computer-readable storage medium is provided, which stores computer-readable instructions. The computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the method for processing the insurance exception list as described above.

[0015] In the scheme implemented by the above-mentioned processing method, device, equipment and medium of the insurance exception list, by obtaining the insurance exception list and the corresponding target historical insurance information, a comprehensive grasp of potential resources is ensured. Then, based on the refined screening rules, the target abnormal insured objects are screened out, which improves the pertinence of subsequent processing. Using the pre-trained classification model, combined with the risk level strategy, the risk category of each target abnormal insured object is accurately divided, and the refinement of risk management is achieved. Then, through the preset circulation rules and risk categories, the circulation path is automatically determined for each target abnormal insured object, avoiding the subjectivity and inefficiency of manual judgment. Finally, the target abnormal insured object and its target historical insurance information are transferred to the corresponding processing terminal, ensuring the effective use and in-depth mining of customer resources, and reducing customer loss due to initial service failure. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the scheme in the present application, a brief introduction is given below to the drawings required for use in the description of the embodiments of the present application. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying any creative work.

[0017] Figure 1 is an exemplary system architecture diagram to which the present application may be applied;

[0018] Figure 2 It is a flowchart of a method for processing an insurance exception list provided by this application;

[0019] Figure 3 It is a structural schematic diagram of a device for processing an insurance exception list provided by the present application;

[0020] Figure 4 It is a structural schematic diagram of a computer device provided by this application. DETAILED DESCRIPTION

[0021] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by technicians in the technical field of the present application; the terms used in the specification of the application herein are only for the purpose of describing specific embodiments and are not intended to limit the present application; the terms "including" and "having" and any variations thereof in the specification and claims of the present application and the above-mentioned drawings are intended to cover non-exclusive inclusions. The terms "first", "second", etc. in the specification and claims of the present application or the above-mentioned drawings are used to distinguish different objects, not to describe a specific order.

[0022] Reference to "embodiments" herein means that a particular feature, structure, or characteristic described in conjunction with the embodiments may be included in at least one embodiment of the present application. The appearance of the phrase in various locations in the specification does not necessarily refer to the same embodiment, nor is it an independent or alternative embodiment that is mutually exclusive with other embodiments. It is explicitly and implicitly understood by those skilled in the art that the embodiments described herein may be combined with other embodiments.

[0023] In order to enable those skilled in the art to better understand the solution of the present application, the technical solution in the embodiments of the present application will be clearly and completely described below in conjunction with the accompanying drawings.

[0024] like Figure 1As shown, the system architecture 100 may include a terminal device 101, a network 102 and a server 103. The terminal device 101 may be a laptop 1011, a tablet computer 1012 or a mobile phone 1013. The network 102 is used to provide a medium for a communication link between the terminal device 101 and the server 103. The network 102 may include various connection types, such as wired, wireless communication links or optical fiber cables.

[0025] The user can use the terminal device 101 to interact with the server 103 through the network 102 to receive or send messages, etc. Various communication client applications can be installed on the terminal device 101, such as web browser applications, shopping applications, search applications, instant messaging tools, email clients, social platform software, etc.

[0026] The terminal device 101 can be any electronic device with a display screen and supporting web browsing. In addition to a laptop computer 1011, a tablet computer 1012 or a mobile phone 1013, the terminal device 101 can also be an e-book reader, an MP3 player (Moving Picture Experts Group Audio Layer III), an MP4 (Moving Picture Experts Group Audio Layer IV), a laptop computer, a desktop computer, etc.

[0027] The server 103 may be a server that provides various services, such as a background server that provides support for a web page displayed on the terminal device 101 .

[0028] It should be noted that the method for processing the insurance exception list provided in the embodiment of the present application is generally executed by a server, and accordingly, the processing device for the insurance exception list is generally set in the server.

[0029] It should be understood that Figure 1 The number of terminal devices, networks and servers in the embodiment is only for illustration. Any number of terminal devices, networks and servers may be provided according to implementation requirements.

[0030] Continue to refer Figure 2 , shows a flow chart of an embodiment of a method for processing an insurance abnormality list according to the present application. The method for processing an insurance abnormality list comprises the following steps:

[0031] Step S201, obtaining an insurance exception list and target historical insurance information corresponding to the insurance exception list, wherein the insurance exception list includes multiple insurance exception objects.

[0032] In this embodiment, the method for processing the insurance exception list is executed on the electronic device (eg Figure 1 The server shown in the figure) can obtain the target historical insurance information through a wired connection or a wireless connection. It should be noted that the above wireless connection method may include but is not limited to 3G / 4G / 5G connection, WiFi connection, Bluetooth connection, WiMAX connection, Zigbee connection, UWB (ultra wideband) connection, and other wireless connection methods currently known or to be developed in the future.

[0033] Among them, the insurance abnormality list refers to a collection of insured objects that record abnormal behaviors such as failure of the initial service or insufficient understanding of non-auto property insurance during the insurance process due to various reasons. It represents abnormal situations in the insurance process and is used to identify and analyze potential service improvement points. For example, a list of customers who refused to purchase non-auto property insurance when they first purchased it or quickly canceled the insurance after purchasing it.

[0034] The target historical insurance information refers to the historical insurance data associated with each insurance abnormality object in the insurance abnormality list, including but not limited to the insurance time, insurance type, premium amount, customer basic information and historical compensation records. It represents the historical behavior characteristics of the insured object and is used to deeply analyze the reasons for insurance abnormalities. For example, a customer has tried to purchase non-auto property insurance many times in the past but failed.

[0035] Among them, multiple insurance anomaly objects refer to multiple independent individuals or entities with abnormal insurance behaviors identified in the insurance anomaly list. These objects may fail in the first service due to insufficient understanding of non-auto property insurance, low trust or other factors. They form the basis of the group that needs further analysis and improvement. For example, a list of 100 customers who failed to purchase non-auto property insurance for the first time is listed in an insurance anomaly list.

[0036] Step S202, based on the preset screening rules and the target historical insurance information, multiple abnormal insurance objects are screened to obtain multiple screened target abnormal insurance objects.

[0037] Among them, screening rules refer to a set of rules based on specific business logic or algorithms, which are used to screen out insured persons who have not committed fraud from the insurance anomaly list. For example, screen out customers who have tried to purchase non-auto property insurance multiple times in the past year but failed, but have committed fraud.

[0038] Among them, multiple target abnormal insured objects refer to a group of insured abnormal objects that have not committed fraud and are screened out from the insurance abnormality list after being processed by the screening rules. For example, under the screening rules, 30 customers with fraudulent behavior are identified from 100 customers who failed to purchase non-auto property insurance for the first time, and they are removed, and 70 customers without fraudulent behavior are retained.

[0039] Step S203, based on the preset risk level strategy and the target historical insurance information, a pre-trained classification model is used to classify multiple target abnormal insurance objects to determine the risk category of each target abnormal insurance object.

[0040] Among them, the risk level strategy refers to a set of standards or methods for assessing the risk level of insured abnormal objects based on their historical insurance information, behavioral characteristics and other factors. These strategies help insurance companies identify the risk levels of different insured objects so that they can take corresponding measures.

[0041] The classification model refers to a model that is based on machine learning or deep learning algorithms and can automatically classify insured abnormal objects after pre-training. The model can output the risk category to which each insured abnormal object belongs based on the input target historical insurance information.

[0042] Among them, risk categories refer to the category labels obtained by the classification model after classifying the insured abnormal objects according to the target historical insurance information and risk level strategy. These categories represent the different risk levels of the insured objects and are used to guide subsequent processing and circulation.

[0043] Step S204, based on the preset circulation rules and risk categories, determine the circulation path of each target abnormal insured object.

[0044] Among them, the transfer rules refer to a set of rules formulated to determine the transfer path of the insured abnormal objects based on their risk categories and other relevant information. These rules ensure that the insured abnormal objects can be accurately and efficiently transferred to the corresponding processing terminals for processing. For example, customers corresponding to the high-risk category are transferred to the senior processing team, customers corresponding to the medium-risk category are transferred to the intermediate processing team, and customers corresponding to the low-risk category are transferred to the general processing team.

[0045] The circulation path refers to the direction and order of the abnormal insured objects in the internal system of the insurance company, determined according to their risk categories and circulation rules. These paths ensure that the abnormal insured objects can be transferred to the corresponding processing department or team in a timely and accurate manner.

[0046] Step S205, each target abnormal insurance object and target historical insurance information are transferred to the corresponding processing terminal according to the flow path.

[0047] The processing terminal refers to the information system or work platform used by the department, team or individual responsible for handling the insured abnormal objects within the insurance company. These terminals receive the insured abnormal objects and their related information according to the flow path and take corresponding processing measures.

[0048] The embodiments of the present application can ensure a comprehensive grasp of potential resources by obtaining an abnormal insurance list and the corresponding target historical insurance information. Then, based on the refined screening rules, the target abnormal insured objects are screened out, thereby improving the pertinence of subsequent processing. By using the pre-trained classification model and combining the risk level strategy, the risk category of each target abnormal insured object is accurately divided, thereby achieving refined risk management. Then, through the preset circulation rules and risk categories, the circulation path is automatically determined for each target abnormal insured object, avoiding the subjectivity and inefficiency of manual judgment. Finally, the target abnormal insured object and its target historical insurance information are transferred to the corresponding processing terminal, ensuring the effective use and in-depth mining of customer resources, and reducing customer loss due to initial service failure.

[0049] In some optional implementations of this embodiment, step 201, obtaining an insurance exception list and target historical insurance information corresponding to the insurance exception list, specifically includes the following steps:

[0050] Obtain historical insurance information of multiple historical insured objects for insurance events, the historical insurance information including insured object information, insurance type information and service time information; extract key information of the insured object information, insurance type information and service time information; construct an insurance event feature vector based on the key information, use a preset clustering algorithm to perform cluster analysis on the insurance event feature vector to obtain a clustering result; determine whether the insurance event is abnormal based on the clustering result; if abnormal, construct an insurance abnormality list based on the historical insured objects corresponding to the insurance event determined to be abnormal; obtain target historical insurance information corresponding to the insurance abnormality list from the historical insurance information.

[0051] Among them, key information refers to the data elements extracted from the insured person information, insurance type information and service time information that are of decisive significance for analyzing insurance behavior. For example, the age, gender, and occupation in the insured person information, the insurance type and amount of insurance in the insurance type information, and the insurance period and renewal frequency in the service time information are all key information. It represents the basic characteristics and insurance preferences of the insured person and is used to accurately describe the outline of insurance behavior.

[0052] Among them, the insurance event feature vector is a multidimensional numerical vector converted from key information by mathematical methods. Each dimension represents a specific key information feature, such as the age of the insured, the selected insurance type, the first insurance time, etc. For example, the feature vector of an insurance event may contain elements such as [age: 30 years old, gender: male, insurance type: non-auto property insurance, first insurance time: January 2022]. The construction of the feature vector is a bridge between the original insurance information and the subsequent clustering analysis algorithm. It enables the insurance event to be represented in a multidimensional space, which is convenient for pattern recognition and classification.

[0053] Among them, the clustering algorithm is an unsupervised learning algorithm, which is used to automatically divide a group of data points (in this embodiment, the insured event feature vector) into several groups or clusters according to a certain similarity metric, so that the data points in the same cluster are highly similar to each other, while the data points between different clusters are less similar.

[0054] In one example, historical insurance information of multiple historical insured objects for insurance events can be obtained from the database, including but not limited to insured object information (such as name, contact information), insurance type information (indicating insurance type, such as insurance A, insurance B, insurance C, etc.) and service time information (such as insurance date, policy effective date, etc.). Subsequently, natural language processing and machine learning techniques can be used to extract key information from the above information, such as the age, gender, specific classification of insurance type, span of service time, etc. of the insured object, to construct an insurance event feature vector. Next, a preset clustering algorithm (such as K-means clustering algorithm) is used to perform cluster analysis on the insurance event feature vector. By calculating the similarity between feature vectors, insurance events with similar insurance behaviors are classified into the same category to obtain clustering results. Then, based on the clustering results, it is determined whether the insurance event is abnormal. In this example, anomalies are defined as events in which insurance purchases fail, which may be caused by incorrect information filling, fraudulent behavior, customer cancellation, or system errors. By comparing the clustering results with the preset normal insurance behavior pattern, clusters that deviate from the normal pattern are identified and determined as abnormal insurance events. For insurance events determined to be abnormal, the corresponding historical insurance object information is further extracted to construct an insurance abnormality list. Subsequently, the target historical insurance information corresponding to these abnormal lists is retrieved from the historical insurance information.

[0055] The embodiment of the present application can ensure the comprehensiveness and accuracy of information by comprehensively collecting historical data containing key elements such as insured object information, insurance type and service time. Subsequently, by extracting the core points of this information, a vector that can accurately reflect the characteristics of the insurance event is constructed, laying a solid foundation for subsequent analysis. The use of a preset clustering algorithm to perform efficient clustering analysis on these feature vectors not only reveals the inherent connections and distribution patterns between insurance events, but also intelligently identifies potential abnormal insurance events based on the differences in clustering results, effectively improving the sensitivity and accuracy of risk identification. Once determined to be abnormal, an insurance anomaly list is immediately constructed, and the relevant target historical insurance information is extracted in a targeted manner, greatly shortening the abnormal event response cycle.

[0056] In some optional implementations of this embodiment, step S202, based on a preset screening rule and target historical insurance information, screens multiple abnormal insurance objects to obtain multiple screened target abnormal insurance objects, specifically including the following steps:

[0057] The target historical insurance information is analyzed according to preset screening rules to determine whether each abnormal insured object has committed fraud; if fraud is committed, the abnormal insured object with fraud is removed from the insurance abnormality list to obtain multiple target abnormal insured objects after screening; if fraud is not committed, the multiple abnormal insured objects are determined as multiple target abnormal insured objects after screening.

[0058] In one example, a list of insurance anomalies in the past month can be extracted from the insurance company's database, which includes 1,000 insurance anomaly objects. At the same time, the historical insurance information of these anomaly objects is obtained, including personal information, insurance records, credit scores, etc. According to the internal risk control strategy of the insurance company, screening rules are formulated, such as credit scores below 600 points, and multiple uses of the same wrong information for insurance, which are considered high-risk features of fraud. Using big data analysis and machine learning algorithms, the historical insurance information of the above 1,000 anomaly objects is deeply analyzed to determine whether each object has fraudulent behavior. For example, it is found that 200 objects have used false addresses for insurance multiple times. According to the analysis results, the 200 anomaly objects with fraudulent behavior are removed from the original list, and 800 target abnormal insurance objects are obtained after screening. These objects mainly fail to be insured due to incorrect information filling or poor credit status, or self-cancellation of insurance, rather than fraudulent behavior.

[0059] The embodiment of the present application can conduct an in-depth analysis of the target historical insurance information through preset screening rules, and accurately determine whether each abnormal insured object has committed fraud. This process effectively avoids the subjectivity and uncertainty of manual judgment. Furthermore, for abnormal insured objects with fraudulent behavior identified, they are automatically removed from the list, thereby ensuring the purity of the list after screening and reducing the potential risks of the insurance company. For abnormal insured objects without fraudulent behavior, they are determined as the target abnormal insured objects after screening, which is convenient for the insurance company to carry out targeted processing and improvement later.

[0060] In some optional implementations of this embodiment, in step S203, based on the preset risk level strategy and the target historical insurance information, a pre-trained classification model is used to classify multiple target abnormal insurance objects, and before determining the risk category of each target abnormal insurance object, the following steps are specifically included:

[0061] Obtain an insured object data set containing known risk category labels; determine the objective function of a preset support vector machine model based on the insured object data set, the objective function aims to maximize the geometric interval from different class samples of the insured object data set to the hyperplane of the support vector machine model, while satisfying the constraint that the function interval is greater than or equal to a preset threshold; use a sequential minimum optimization algorithm to solve the optimal normal vector and intercept parameters in the support vector machine model; construct a classification decision function based on the optimal normal vector and intercept parameters; determine a pre-trained classification model based on the classification decision function and the support vector machine model.

[0062] Among them, the insured object dataset is extracted from the insurance company database, which contains the basic information of the insured objects and a collection of their known risk category labels. These data are derived from actual insurance records, and represent the characteristics of the insured objects (such as age, occupation, historical claims records, etc.) and their risk categories (such as high risk, medium risk, and low risk). This dataset is used to train the support vector machine model to achieve automatic classification of the risk categories of the insured objects. For example, a dataset containing 1,000 insured objects and their risk category labels can be used to train the model to identify insured objects of different risk levels.

[0063] Among them, the support vector machine model is a classification model based on statistical learning theory, which aims to separate data points of different categories as much as possible by finding a hyperplane. The model learns from the insured object data set and characterizes the complex relationship between the insured object characteristics and the target risk category.

[0064] Among them, the objective function is the function that needs to be minimized during the training process of the support vector machine model, which aims to maximize the geometric interval from different class samples of the insured object data set to the hyperplane of the support vector machine model. This function characterizes the quality of the model classification performance. By optimizing the objective function, the optimal model parameters can be found. The optimization of the objective function helps to improve the accuracy of risk classification. For example, the objective function may aim to maximize the distance from high-risk and low-risk insured objects to the hyperplane.

[0065] Among them, the hyperplane is the core concept in the support vector machine model. It is a plane that separates different categories of the insured object data set (a hyperplane in multidimensional space). The position and direction of the hyperplane are determined by the support vector, which represents the optimal segmentation boundary between different risk categories.

[0066] Among them, the function interval is a measure used in the support vector machine model to measure the distance from the sample point to the hyperplane, which represents the degree to which the sample point is correctly classified. The larger the function interval, the higher the confidence that the sample point is correctly classified.

[0067] The preset threshold is a parameter set during the training process of the support vector machine model to determine whether the sample point is correctly classified. Sample points whose function interval is greater than or equal to the preset threshold are considered to be correctly classified, otherwise they are considered to be incorrectly classified.

[0068] Among them, the constraints are the conditions that must be met during the training process of the support vector machine model. It ensures that the model can maintain the correct classification of sample points while optimizing the objective function.

[0069] Among them, the sequential minimum optimization algorithm is an iterative algorithm for solving the optimal normal vector and intercept parameters in the support vector machine model. It gradually approaches the optimal solution by selecting two sample points for optimization each time. It is used to efficiently train the support vector machine model and accurately classify the risk category of the insured object. For example, the sequential minimum optimization algorithm can find the optimal model parameters in a short time, so that the support vector machine model can accurately distinguish between high-risk and low-risk insured objects.

[0070] The optimal normal vector is the vector found during the training of the support vector machine model, which determines the direction and position of the hyperplane. The optimal normal vector represents the optimal relationship between the insured object characteristics and the target risk category, and is a key parameter for the model classification performance.

[0071] The intercept parameter is a constant term in the support vector machine model, which together with the optimal normal vector determines the position of the hyperplane. The intercept parameter characterizes the offset of the model during the classification process, and the setting of the intercept parameter can affect the sensitivity and specificity of risk classification.

[0072] The classification decision function is a function built based on the optimal normal vector and intercept parameter of the support vector machine model, which is used to determine the risk category to which the insured object belongs. The function inputs the characteristics of the insured object into the model and outputs a label representing the risk category.

[0073] In one example, a data set of insured objects containing known risk category labels can be obtained from an insurance company database. The data set may include basic information of the insured (such as age, occupation, income, etc.), historical claims records, and credit scores, and each sample is labeled as a high-risk or low-risk category. Based on the above data set, the objective function of the support vector machine model is designed. The objective function aims to maximize the geometric interval from samples of different risk categories in the insured object data set to the hyperplane of the support vector machine model, while ensuring that the function interval is not less than a preset threshold (e.g., the preset threshold is set to 1), thereby enhancing the classification robustness of the model. The sequential minimum optimization algorithm is used to iteratively update the Lagrange multiplier corresponding to the support vector to solve the optimal normal vector (i.e., weight vector) and intercept parameter in the support vector machine model. This process effectively reduces the computational complexity and improves the model training efficiency. Based on the obtained optimal normal vector and intercept parameter, a classification decision function is constructed to determine whether the new insured object belongs to the high-risk or low-risk category. The trained support vector machine model is combined with the classification decision function to form a pre-trained classification model. This model can accurately identify the risk level of the insured object.

[0074] The embodiment of the present application can ensure the accuracy and pertinence of the training data by obtaining a data set of insured objects containing known risk category labels. Subsequently, the preset support vector machine model is used to construct the objective function, which aims to maximize the geometric interval of samples of different categories to the hyperplane, and set a preset threshold of the function interval, which effectively enhances the classification boundary robustness of the model. The sequential minimum optimization algorithm is used to solve the optimal normal vector and intercept parameters. This process not only optimizes the computational efficiency, but also ensures the accuracy of the model parameters. The classification decision function constructed based on these parameters can realize the accurate division of the risk category of the insured object. Finally, combined with the support vector machine model, a pre-trained classification model is formed, which shows a high degree of accuracy and generalization ability in insurance risk assessment.

[0075] In some optional implementations of this embodiment, step S203, based on a preset risk level strategy and target historical insurance information, uses a pre-trained classification model to classify multiple target abnormal insurance objects and determine the risk category of each target abnormal insurance object, specifically including the following steps:

[0076] Attribute fields related to risk assessment are extracted from the target historical insurance information; based on the attribute fields, an abnormal insured object feature vector is constructed for each target abnormal insured object; according to a preset risk level strategy, the abnormal insured object feature vector is analyzed, and the risk score of each target abnormal insured object is calculated to form a risk score vector; the risk score vector is input into the classification model, and the abnormal insured objects are classified and predicted through the classification decision function of the classification model to determine the risk category of each target abnormal insured object.

[0077] Among them, attribute fields refer to data elements or information identifiers extracted from the target's historical insurance information and directly related to risk assessment. They characterize the risk characteristics and behavior patterns of the insured object.

[0078] Among them, the abnormal insured object feature vector is a vector representation formed by quantifying the multi-dimensional risk characteristics of each target abnormal insured object based on the extracted attribute fields.

[0079] Among them, the risk score refers to the quantitative assessment value obtained by analyzing the abnormal insured object feature vector of each target abnormal insured object according to the preset risk level strategy.

[0080] The risk score vector is a set of vectors formed by arranging the risk scores of all target abnormal insured objects in order. Each element represents the risk score of a target abnormal insured object, and the entire vector reflects the risk distribution characteristics of the entire group of abnormal insured objects.

[0081] In one example, an abnormal insurance list can be obtained, which includes multiple insured persons who have recently shown abnormal behaviors in the process of home property insurance, such as frequent withdrawal of insurance, refusal of claims, etc. Correspondingly, the target historical insurance information of these abnormal insured persons is collected, including but not limited to the basic information of the insured (age, gender, occupation), family property status, historical compensation records, insurance purchase frequency, etc. Based on the preset screening rules, multiple abnormal insured persons are screened to obtain a set of screened target abnormal insured objects. Attribute fields directly related to risk assessment are extracted from the target historical insurance information, such as the insured's age, occupational stability, historical compensation amount, number of compensations, and insurance product type diversity. Based on these attribute fields, an abnormal insured object feature vector is constructed for each target abnormal insured object, and each element in the vector represents the quantized value of an attribute field. According to the preset risk level strategy, the feature vector is analyzed to calculate the risk score of each target abnormal insured object. The risk score comprehensively considers factors such as the insured's compensation history, property status changes, and purchase behavior. The risk scores of all target abnormal insured objects are arranged in order to form a risk score vector. The risk score vector is input into a pre-trained classification model (e.g., pre-trained support vector machine, random forest, etc.), which has learned the mapping relationship between different risk categories and risk scores based on a large amount of historical data. Through the classification decision function of the classification model, the abnormal insured objects are classified and predicted to determine the risk category of each target abnormal insured object, such as high risk, medium risk, low risk, etc.

[0082] The embodiment of the present application can construct a comprehensive feature vector for each target abnormal insured object by accurately extracting the risk assessment attribute field in the target historical insurance information, thereby ensuring the comprehensiveness and accuracy of the risk assessment. Subsequently, the feature vector is deeply analyzed based on the preset risk level strategy, and the risk score is calculated to form an intuitive risk score vector, which provides a quantitative basis for risk assessment. The risk score vector is input into a trained classification model, and with the help of the classification decision function of the model, efficient classification prediction of abnormal insured objects is achieved, and the risk category of each target abnormal insured object is accurately divided.

[0083] In some optional implementations of this embodiment, step S204, based on the preset circulation rules and risk categories, determines the circulation path of each target abnormal insurance object, specifically including the following steps:

[0084] Obtain preset circulation rules, which include multiple risk category conditions and circulation paths corresponding to each risk category condition; match the risk category of each target abnormal insured object with each risk category condition to obtain a matching result; based on the matching result, determine the circulation paths corresponding to the risk category conditions that match the risk category as the circulation paths of multiple target abnormal insured objects.

[0085] Among them, risk category conditions refer to the specific judgment basis or standards used to distinguish and identify different risk categories in the preset circulation rules.

[0086] In one example, the insurance company presets a set of circulation rules based on historical data and business experience. The rules include multiple risk category conditions and the circulation path corresponding to each condition. Based on the preset risk level strategy and pre-trained classification model, the risk category of each target abnormal insured object has been determined, such as "high risk", "medium risk", "low risk", etc. Subsequently, the risk category of each target abnormal insured object is matched one by one with the risk category conditions in the preset circulation rules to find the most suitable risk category condition. Once a matching risk category condition is found, the circulation path corresponding to the condition is determined. For example, if a target abnormal insured object is classified as "high risk" and its risk characteristics match the condition of "the claim ratio exceeds a certain threshold", it will be transferred to the processing terminal of the "high risk processing team".

[0087] The embodiment of the present application can realize intelligent determination of the circulation path by accurately matching the risk category of each target abnormal insured object with the risk category conditions in the preset circulation rules. Make full use of the multiple risk category conditions and corresponding circulation paths defined in the circulation rules to ensure that each target abnormal insured object can be accurately directed to the processing flow that best suits its risk characteristics. This intelligent matching process not only improves circulation efficiency and reduces the need for manual intervention, but also ensures the accuracy and pertinence of the circulation path.

[0088] In some optional implementations of this embodiment, in step S205, after each target abnormal insurance object and target historical insurance information are transferred to the corresponding processing terminal according to the flow path, the following steps are specifically included:

[0089] Receive feedback data sent by the processing terminal, where the feedback data is the service result data returned by the processing terminal for providing in-depth services to each target abnormally insured object based on the target historical insurance information; use an association rule mining algorithm to perform association mining on the feedback data and the flow rules to obtain the association rules between the feedback data and the flow rules; generate flow rule optimization suggestions based on the association rules, adjust the flow rules based on the flow rule optimization suggestions, and obtain optimized flow rules.

[0090] Among them, feedback data refers to the service result data returned by the processing terminal after providing in-depth services to each target abnormal insured object based on the received target historical insurance information. It represents the analysis, evaluation or processing results of the abnormal insured object by the processing terminal, and is used to evaluate the effectiveness of in-depth services and the effect of abnormal processing.

[0091] Among them, deep service refers to a series of meticulous and in-depth analysis and processing processes performed by the processing terminal on the target abnormal insured objects after screening and classification based on their target historical insurance information. For example, deep service may include detailed underwriting investigation, risk assessment report preparation, personalized insurance plan design, etc.

[0092] Among them, the association rule mining algorithm is used to analyze the relationship between feedback data and circulation rules to reveal the potential connection between the two. For example, the algorithm may find that a certain type of abnormal insured object is more likely to receive positive feedback under a specific circulation path, thereby guiding the adjustment of circulation rules.

[0093] Among them, rule optimization suggestions refer to suggestions for improving or adjusting existing circulation rules based on the results of association mining. They represent the identification and improvement direction of the shortcomings of existing rules. Rule optimization suggestions are used to guide the revision and improvement of circulation rules to improve the efficiency and accuracy of business processing.

[0094] In one example, when processing an abnormal insurance list, the insurance company first obtains a list of multiple abnormal insured objects and their corresponding target historical insurance information. After processing according to the preset screening rules, the screened target abnormal insured objects are obtained. These objects are then further classified by the classification model according to the risk level strategy to determine their respective risk categories. Based on the risk category and the preset flow rules, each target abnormal insured object is assigned a corresponding flow path and transferred to the corresponding processing terminal. After receiving the target abnormal insured object and its historical insurance information, the processing terminal, such as the underwriting department or risk assessment team, provides in-depth services. After the service is completed, the processing terminal returns feedback data, which records the processing results of each target abnormal insured object in detail. After collecting a large amount of feedback data, the insurance company uses an association rule mining algorithm to associate these data with the flow rules. The algorithm analyzes the characteristics of the feedback data returned by the processing terminal under different flow paths, reveals the potential connection between the flow rules and the processing results, that is, determines the association rules between the flow rules and the processing results. Based on the association rules, the insurance company generates flow rule optimization suggestions. These suggestions may include adjusting the circulation path of objects of specific risk categories, increasing or reducing certain processing links, optimizing resource allocation of processing terminals, etc. Subsequently, the insurance company adjusted the circulation rules based on these suggestions and obtained optimized circulation rules.

[0095] The embodiment of the present application can obtain the actual effect of business processing in real time by receiving and processing the feedback data returned by the terminal after providing in-depth services to the target abnormal insured object based on the target historical insurance information. The association rule mining algorithm is used to deeply mine the potential connection between the feedback data and the existing circulation rules, which not only makes full use of the rich business data resources, but also combines the common knowledge about data processing and rule optimization in the field, effectively improving the accuracy and efficiency of data analysis. Based on the association rules obtained by mining, the generated circulation rule optimization suggestions are targeted and operational, and can accurately point out the problems and improvement directions in the existing circulation rules. By making targeted adjustments to the circulation rules, the optimized circulation rules are more in line with the actual business needs, effectively improving the automation level and response speed of business processing, and reducing the complexity and cost of human intervention.

[0096] It should be emphasized that in order to further ensure the above-mentioned insurance exception list, target historical insurance information and historical insurance information, the above-mentioned insurance exception list, target historical insurance information and historical insurance information can also be stored in a node of a blockchain.

[0097] The blockchain referred to in this application is a new application model of computer technologies such as distributed data storage, peer-to-peer transmission, consensus mechanism, encryption algorithm, etc. Blockchain is essentially a decentralized database, a string of data blocks generated by cryptographic methods. Each data block contains a batch of network transaction information, which is used to verify the validity of its information (anti-counterfeiting) and generate the next block. Blockchain can include the underlying blockchain platform, platform product service layer, and application service layer.

[0098] The embodiments of the present application can build and optimize related models and networks based on artificial intelligence technology, such as classification models, support vector machine models, etc. Among them, artificial intelligence (AI) models are the crystallization of theory and practice that simulates the decision-making process of human intelligence through algorithms and data analysis to solve complex problems, predict future trends, or realize automation tasks. These models use a large amount of historical data and real-time information, and are trained and optimized through a specific algorithm framework to achieve efficient, accurate and reliable performance.

[0099] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through computer-readable instructions, and the computer-readable instructions can be stored in a computer-readable storage medium. When the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the aforementioned storage medium can be a non-volatile storage medium such as a disk, an optical disk, a read-only memory (ROM), or a random access memory (RAM).

[0100] It should be understood that, although the steps in the flowchart of the accompanying drawings are displayed in sequence as indicated by the arrows, these steps are not necessarily executed in sequence in the order indicated by the arrows. Unless otherwise specified herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least a part of the steps in the flowchart of the accompanying drawings may include multiple sub-steps or multiple stages, and these sub-steps or stages are not necessarily executed at the same time, but can be executed at different times, and their execution order is not necessarily sequential, but can be executed in turn or alternately with other steps or at least a part of the sub-steps or stages of other steps.

[0101] Further references Figure 3 , as a response to the above Figure 2 In order to realize the method shown in the figure, the present application provides an embodiment of a device for processing an insurance exception list, and the device embodiment is similar to Figure 2 Corresponding to the method embodiment shown, the device can be specifically applied to various electronic devices.

[0102] like Figure 3 As shown, the device 400 for processing the insurance abnormality list of this embodiment includes: a first acquisition module 401, a screening module 402, a classification module 403, a path determination module 404 and a circulation module 405. Among them:

[0103] The first acquisition module 401 is used to acquire an insurance exception list and target historical insurance information corresponding to the insurance exception list, wherein the insurance exception list includes multiple insurance exception objects;

[0104] A screening module 402 is used to screen multiple abnormal insurance objects based on preset screening rules and target historical insurance information to obtain multiple screened target abnormal insurance objects;

[0105] The classification module 403 is used to classify multiple target abnormal insurance objects based on the preset risk level strategy and the target historical insurance information, using a pre-trained classification model, and determine the risk category of each target abnormal insurance object;

[0106] A path determination module 404 is used to determine the flow path of each target abnormal insurance object based on preset flow rules and risk categories;

[0107] The transfer module 405 is used to transfer each target abnormal insurance object and target historical insurance information to the corresponding processing terminal according to the transfer path.

[0108] The embodiments of the present application can ensure a comprehensive grasp of potential resources by obtaining an abnormal insurance list and the corresponding target historical insurance information. Then, based on the refined screening rules, the target abnormal insured objects are screened out, thereby improving the pertinence of subsequent processing. By using the pre-trained classification model and combining the risk level strategy, the risk category of each target abnormal insured object is accurately divided, thereby achieving refined risk management. Then, through the preset circulation rules and risk categories, the circulation path is automatically determined for each target abnormal insured object, avoiding the subjectivity and inefficiency of manual judgment. Finally, the target abnormal insured object and its target historical insurance information are transferred to the corresponding processing terminal, ensuring the effective use and in-depth mining of customer resources, and reducing customer loss due to initial service failure.

[0109] In one embodiment, the first acquisition module 401 includes:

[0110] The first acquisition submodule is used to acquire historical insurance information of multiple historical insured objects for insurance events, where the historical insurance information includes insured object information, insurance type information, and service time information;

[0111] The information extraction submodule is used to extract key information such as the insured person information, insurance type information and service time information;

[0112] The cluster analysis submodule is used to construct an insurance event feature vector based on key information, and to perform cluster analysis on the insurance event feature vector using a preset clustering algorithm to obtain a clustering result;

[0113] A judgment submodule is used to judge whether the insurance event is abnormal based on the clustering results;

[0114] The determination submodule is used to construct an insurance abnormality list based on the historical insurance objects corresponding to the insurance events determined to be abnormal if there is an abnormality;

[0115] The second acquisition submodule is used to obtain target historical insurance information corresponding to the insurance exception list from the historical insurance information.

[0116] The embodiment of the present application can ensure the comprehensiveness and accuracy of information by comprehensively collecting historical data containing key elements such as insured object information, insurance type and service time. Subsequently, by extracting the core points of this information, a vector that can accurately reflect the characteristics of the insurance event is constructed, laying a solid foundation for subsequent analysis. The use of a preset clustering algorithm to perform efficient clustering analysis on these feature vectors not only reveals the inherent connections and distribution patterns between insurance events, but also intelligently identifies potential abnormal insurance events based on the differences in clustering results, effectively improving the sensitivity and accuracy of risk identification. Once determined to be abnormal, an insurance anomaly list is immediately constructed, and the relevant target historical insurance information is extracted in a targeted manner, greatly shortening the abnormal event response cycle.

[0117] In one embodiment, the screening module 402 includes:

[0118] The first analysis submodule is used to analyze the target historical insurance information according to the preset screening rules to determine whether each abnormal insurance object has fraudulent behavior;

[0119] The removal submodule is used to remove the abnormal insured person with fraudulent behavior from the insurance abnormality list if there is fraudulent behavior, so as to obtain multiple target abnormal insured persons after screening;

[0120] The first determination submodule is used to determine the multiple abnormal insured persons as the multiple screened target abnormal insured persons if there is no fraudulent behavior.

[0121] The embodiment of the present application can conduct an in-depth analysis of the target historical insurance information through preset screening rules, and accurately determine whether each abnormal insured object has committed fraud. This process effectively avoids the subjectivity and uncertainty of manual judgment. Furthermore, for abnormal insured objects with fraudulent behavior identified, they are automatically removed from the list, thereby ensuring the purity of the list after screening and reducing the potential risks of the insurance company. For abnormal insured objects without fraudulent behavior, they are determined as the target abnormal insured objects after screening, which is convenient for the insurance company to carry out targeted processing and improvement later.

[0122] In one embodiment, the classification module 403 includes:

[0123] A field extraction submodule is used to extract attribute fields related to risk assessment from the target historical insurance information;

[0124] A construction submodule is used to construct an abnormal insurance object feature vector for each target abnormal insurance object based on the attribute field;

[0125] The second analysis submodule is used to analyze the feature vector of the abnormal insured object according to the preset risk level strategy, calculate the risk score of each target abnormal insured object, and form a risk score vector;

[0126] The prediction submodule is used to input the risk score vector into the classification model, perform classification prediction on the abnormal insured objects through the classification decision function of the classification model, and determine the risk category of each target abnormal insured object.

[0127] The embodiment of the present application can construct a comprehensive feature vector for each target abnormal insured object by accurately extracting the risk assessment attribute field in the target historical insurance information, thereby ensuring the comprehensiveness and accuracy of the risk assessment. Subsequently, the feature vector is deeply analyzed based on the preset risk level strategy, and the risk score is calculated to form an intuitive risk score vector, which provides a quantitative basis for risk assessment. The risk score vector is input into a trained classification model, and with the help of the classification decision function of the model, efficient classification prediction of abnormal insured objects is achieved, and the risk category of each target abnormal insured object is accurately divided.

[0128] In one embodiment, the path determination module 404 includes:

[0129] The third acquisition submodule is used to acquire a preset flow rule, where the flow rule includes a plurality of risk category conditions and a flow path corresponding to each risk category condition;

[0130] A matching submodule is used to match the risk category of each target abnormal insurance object with each risk category condition to obtain a matching result;

[0131] The second determination submodule is used to determine the flow paths corresponding to the risk category conditions matching the risk categories as the flow paths of multiple target abnormal insured objects based on the matching results.

[0132] The embodiment of the present application can realize intelligent determination of the circulation path by accurately matching the risk category of each target abnormal insured object with the risk category conditions in the preset circulation rules. Make full use of the multiple risk category conditions and corresponding circulation paths defined in the circulation rules to ensure that each target abnormal insured object can be accurately directed to the processing flow that best suits its risk characteristics. This intelligent matching process not only improves circulation efficiency and reduces the need for manual intervention, but also ensures the accuracy and pertinence of the circulation path.

[0133] In one embodiment, the device 400 for processing the insurance exception list further includes:

[0134] The second acquisition module is used to acquire an insured object data set containing known risk category labels;

[0135] A function determination module is used to determine the objective function of a preset support vector machine model according to the insured object data set, wherein the objective function aims to maximize the geometric interval from different class samples of the insured object data set to the hyperplane of the support vector machine model, while satisfying the constraint condition that the function interval is greater than or equal to a preset threshold;

[0136] A solution module, used for solving the optimal normal vector and intercept parameters in the support vector machine model by using a sequential minimum optimization algorithm;

[0137] A construction module for constructing a classification decision function based on an optimal normal vector and an intercept parameter;

[0138] The model determination module is used to determine the pre-trained classification model based on the classification decision function and the support vector machine model.

[0139] The embodiment of the present application can ensure the accuracy and pertinence of the training data by obtaining a data set of insured objects containing known risk category labels. Subsequently, the preset support vector machine model is used to construct the objective function, which aims to maximize the geometric interval of samples of different categories to the hyperplane, and set a preset threshold of the function interval, which effectively enhances the classification boundary robustness of the model. The sequential minimum optimization algorithm is used to solve the optimal normal vector and intercept parameters. This process not only optimizes the computational efficiency, but also ensures the accuracy of the model parameters. The classification decision function constructed based on these parameters can realize the accurate division of the risk category of the insured object. Finally, combined with the support vector machine model, a pre-trained classification model is formed, which shows a high degree of accuracy and generalization ability in insurance risk assessment.

[0140] In one embodiment, the device 400 for processing the insurance exception list further includes:

[0141] A receiving module, used to receive feedback data sent by a processing terminal, wherein the feedback data is service result data returned by the processing terminal for performing in-depth services on each target abnormal insurance object based on the target historical insurance information;

[0142] A mining module is used to use an association rule mining algorithm to perform association mining on the feedback data and the flow rules to obtain the association rules between the feedback data and the flow rules;

[0143] The adjustment module is used to generate flow rule optimization suggestions based on the association rules, and adjust the flow rules based on the flow rule optimization suggestions to obtain optimized flow rules.

[0144] The embodiment of the present application can obtain the actual effect of business processing in real time by receiving and processing the feedback data returned by the terminal after providing in-depth services to the target abnormal insured object based on the target historical insurance information. The association rule mining algorithm is used to deeply mine the potential connection between the feedback data and the existing circulation rules, which not only makes full use of the rich business data resources, but also combines the common knowledge about data processing and rule optimization in the field, effectively improving the accuracy and efficiency of data analysis. Based on the association rules obtained by mining, the generated circulation rule optimization suggestions are targeted and operational, and can accurately point out the problems and improvement directions in the existing circulation rules. By making targeted adjustments to the circulation rules, the optimized circulation rules are more in line with the actual business needs, effectively improving the automation level and response speed of business processing, and reducing the complexity and cost of human intervention.

[0145] To solve the above technical problems, the present application also provides a computer device. Figure 4 , Figure 4 This is a basic structural block diagram of the computer device in this embodiment.

[0146] The computer device 6 includes a memory 61, a processor 62, and a network interface 63 that are interconnected and communicated through a system bus. It should be noted that the figure only shows a computer device 6 having a memory 61, a processor 62, and a network interface 63, but it should be understood that it is not required to implement all the components shown, and more or fewer components can be implemented instead. Among them, those skilled in the art can understand that the computer device here is a device that can automatically perform numerical calculations and / or information processing according to pre-set or stored instructions, and its hardware includes but is not limited to microprocessors, application specific integrated circuits (ASIC), programmable gate arrays (FPGA), digital signal processors (DSP), embedded devices, etc.

[0147] Computer devices can be computing devices such as desktop computers, notebooks, PDAs, and cloud servers. Computer devices can interact with users through keyboards, mice, remote controls, touch pads, or voice control devices.

[0148] The memory 61 includes at least one type of readable storage medium, and the readable storage medium includes flash memory, hard disk, multimedia card, card-type memory (for example, SD or DX memory, etc.), random access memory (RAM), static random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, disk, optical disk, etc. In some embodiments, the memory 61 can be an internal storage unit of the computer device 6, such as a hard disk or memory of the computer device 6. In other embodiments, the memory 61 can also be an external storage device of the computer device 6, such as a plug-in hard disk equipped on the computer device 6, a smart memory card (SmartMedia Card, SMC), a secure digital (Secure Digital, SD) card, a flash card (Flash Card), etc. Of course, the memory 61 can also include both the internal storage unit of the computer device 6 and its external storage device. In this embodiment, the memory 61 is generally used to store the operating system and various application software installed on the computer device 6, such as computer-readable instructions of the method for processing the insurance abnormality list, etc. In addition, the memory 61 can also be used to temporarily store various types of data that have been output or are to be output.

[0149] The processor 62 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chips in some embodiments. The processor 62 is generally used to control the overall operation of the computer device 6. In this embodiment, the processor 62 is used to run computer-readable instructions stored in the memory 61 or process data, such as computer-readable instructions for running a method for processing an insurance exception list.

[0150] The network interface 63 may include a wireless network interface or a wired network interface, and the network interface 63 is generally used to establish a communication connection between the computer device 6 and other electronic devices.

[0151] The embodiments of the present application can ensure a comprehensive grasp of potential resources by obtaining an abnormal insurance list and the corresponding target historical insurance information. Then, based on the refined screening rules, the target abnormal insured objects are screened out, thereby improving the pertinence of subsequent processing. By using the pre-trained classification model and combining the risk level strategy, the risk category of each target abnormal insured object is accurately divided, thereby achieving refined risk management. Then, through the preset circulation rules and risk categories, the circulation path is automatically determined for each target abnormal insured object, avoiding the subjectivity and inefficiency of manual judgment. Finally, the target abnormal insured object and its target historical insurance information are transferred to the corresponding processing terminal, ensuring the effective use and in-depth mining of customer resources, and reducing customer loss due to initial service failure.

[0152] The present application also provides another implementation, namely, providing a computer-readable storage medium, which stores computer-readable instructions, and the computer-readable instructions can be executed by at least one processor to enable the at least one processor to perform the steps of the method for processing the insurance exception list as described above.

[0153] The embodiments of the present application can ensure a comprehensive grasp of potential resources by obtaining an abnormal insurance list and the corresponding target historical insurance information. Then, based on the refined screening rules, the target abnormal insured objects are screened out, thereby improving the pertinence of subsequent processing. By using the pre-trained classification model and combining the risk level strategy, the risk category of each target abnormal insured object is accurately divided, thereby achieving refined risk management. Then, through the preset circulation rules and risk categories, the circulation path is automatically determined for each target abnormal insured object, avoiding the subjectivity and inefficiency of manual judgment. Finally, the target abnormal insured object and its target historical insurance information are transferred to the corresponding processing terminal, ensuring the effective use and in-depth mining of customer resources, and reducing customer loss due to initial service failure.

[0154] Through the description of the above implementation methods, those skilled in the art can clearly understand that the above-mentioned embodiment methods can be implemented by means of software plus a necessary general hardware platform, and of course by hardware, but in many cases the former is a better implementation method. Based on such an understanding, the technical solution of the present application, or the part that contributes to the prior art, can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, a disk, or an optical disk), and includes a number of instructions for a terminal device (which can be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods of each embodiment of the present application.

[0155] Obviously, the embodiments described above are only a part of the embodiments of the present application, rather than all the embodiments. The preferred embodiments of the present application are given in the accompanying drawings, but they do not limit the patent scope of the present application. The present application can be implemented in many different forms. On the contrary, the purpose of providing these embodiments is to make the understanding of the disclosure of the present application more thorough and comprehensive. Although the present application is described in detail with reference to the aforementioned embodiments, for those skilled in the art, it is still possible to modify the technical solutions recorded in the aforementioned specific embodiments, or to replace some of the technical features therein with equivalents. All equivalent structures made using the contents of the specification and drawings of this application, directly or indirectly used in other related technical fields, are similarly within the scope of patent protection of this application. The non-company software tools or components that appear in the embodiments of this application are merely examples and do not represent actual use.

Claims

1. A method for processing an insurance exception list, characterized in that: The steps include: Acquire an insurance exception list and target historical insurance information corresponding to the insurance exception list, wherein the insurance exception list includes multiple insurance exception objects; Based on the preset screening rules and the target historical insurance information, the multiple abnormal insurance objects are screened to obtain multiple screened target abnormal insurance objects; Based on the preset risk level strategy and the target historical insurance information, a pre-trained classification model is used to classify the multiple target abnormal insurance objects, and determine the risk category of each target abnormal insurance object; Based on the preset circulation rules and the risk category, determining the circulation path of each target abnormal insurance object; Each target abnormal insurance object and the target historical insurance information are transferred to the corresponding processing terminal according to the flow path.

2. The method according to claim 1, characterized in that The step of obtaining the insurance exception list and the target historical insurance information corresponding to the insurance exception list specifically includes: Acquire historical insurance information of multiple historical insured objects for insurance events, wherein the historical insurance information includes insured object information, insurance type information, and service time information; Extract key information of the insured object information, the insurance type information and the service time information; Constructing an insurance event feature vector based on the key information, and using a preset clustering algorithm to perform cluster analysis on the insurance event feature vector to obtain a clustering result; Determining whether the insurance event is abnormal based on the clustering result; If there is an abnormality, a list of insurance abnormalities is constructed based on the historical insured objects corresponding to the insurance event determined to be abnormal; The target historical insurance information corresponding to the insurance exception list is obtained from the historical insurance information.

3. The method according to claim 1, characterized in that The step of screening the plurality of abnormal insured objects based on the preset screening rules and the target historical insurance information to obtain the screened plurality of target abnormal insured objects specifically includes: Analyze the target historical insurance information according to the preset screening rules to determine whether each abnormal insurance object has fraudulent behavior; If there is fraudulent behavior, the abnormal insured person with fraudulent behavior is removed from the insurance abnormality list to obtain a plurality of target abnormal insured persons after screening; If there is no fraud, the multiple abnormal insured persons are determined as the multiple target abnormal insured persons after screening.

4. The method according to claim 1, characterized in that: Before the step of classifying the plurality of target abnormal insured objects based on the preset risk level strategy and the target historical insurance information using a pre-trained classification model to determine the risk category of each target abnormal insured object, the method further includes: Obtain a dataset of insured persons with known risk category labels; Determine, according to the insured object data set, an objective function of a preset support vector machine model, wherein the objective function aims to maximize the geometric interval from different class samples of the insured object data set to the hyperplane of the support vector machine model, while satisfying a constraint condition that the function interval is greater than or equal to a preset threshold; Using a sequential minimum optimization algorithm to solve the optimal normal vector and intercept parameters in the support vector machine model; Based on the optimal normal vector and the intercept parameter, construct a classification decision function; Based on the classification decision function and the support vector machine model, a pre-trained classification model is determined.

5. The method according to claim 4, characterized in that The step of classifying the plurality of target abnormal insured objects based on the preset risk level strategy and the target historical insurance information using a pre-trained classification model and determining the risk category of each target abnormal insured object specifically includes: Extracting attribute fields related to risk assessment from the target historical insurance information; Based on the attribute field, construct an abnormal insured object feature vector for each target abnormal insured object; According to a preset risk level strategy, the feature vector of the abnormal insured object is analyzed, and the risk score of each target abnormal insured object is calculated to form a risk score vector; The risk score vector is input into the classification model, and the abnormal insured objects are classified and predicted through the classification decision function of the classification model to determine the risk category of each target abnormal insured object.

6. The method according to claim 1, characterized in that The step of determining the circulation path of each target abnormal insured object based on the preset circulation rule and the risk category specifically includes: Obtaining a preset flow rule, wherein the flow rule includes a plurality of risk category conditions and a flow path corresponding to each risk category condition; Matching the risk category of each target abnormal insurance object with each risk category condition to obtain a matching result; Based on the matching results, the flow paths corresponding to the risk category conditions that match the risk category are respectively determined as the flow paths of the multiple target abnormal insured objects.

7. The method according to claim 1, characterized in that After the step of transferring each target abnormal insurance object and the target historical insurance information to the corresponding processing terminal according to the flow path, the method further includes: receiving feedback data sent by the processing terminal, wherein the feedback data is service result data returned by the processing terminal for performing in-depth services on each target abnormal insurance object based on the target historical insurance information; Using an association rule mining algorithm to perform association mining on the feedback data and the flow rule to obtain an association rule between the feedback data and the flow rule; Generate a flow rule optimization suggestion based on the association rule, and adjust the flow rule based on the flow rule optimization suggestion to obtain an optimized flow rule.

8. A device for processing an insurance abnormality list, characterized in that: include: A first acquisition module is used to acquire an insurance exception list and target historical insurance information corresponding to the insurance exception list, wherein the insurance exception list includes multiple insurance exception objects; A screening module, used for screening the plurality of abnormal insured objects based on a preset screening rule and the target historical insurance information, to obtain a plurality of screened target abnormal insured objects; A classification module, for classifying the plurality of target abnormal insured objects based on a preset risk level strategy and the target historical insurance information, using a pre-trained classification model, and determining the risk category of each target abnormal insured object; A path determination module, used to determine the flow path of each target abnormal insurance object based on a preset flow rule and the risk category; The transfer module is used to transfer each target abnormal insurance object and the target historical insurance information to the corresponding processing terminal according to the transfer path.

9. A computer device, characterized in that: It includes a memory and a processor, wherein the memory stores computer-readable instructions, and when the processor executes the computer-readable instructions, it implements the steps of the method for processing the insurance exception list as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer-readable instructions, which, when executed by a processor, implement the steps of the method for processing an insurance exception list as described in any one of claims 1 to 7.

Images