Car insurance claim settlement method, car insurance claim settlement device, electronic equipment and storage medium

Abstract

The embodiment of the invention provides a car insurance claim settlement method, a car insurance claim settlement device, electronic equipment and a storage medium, and belongs to the technical field of financial science and technology. The method comprises the following steps: performing claim feature extraction on a vehicle insurance policy, a first historical driving behavior in a first driving period, a first historical position and a first historical driving environment to obtain an initial claim hidden state, claim settlement feature extraction is carried out on the vehicle insurance policy, the second historical driving behavior of the second driving period, the second historical position, the second historical driving environment and the initial claim settlement hidden state to obtain a target claim settlement hidden state, the claim settlement category of the target vehicle is determined according to the target claim settlement hidden state, and the claim settlement category is used for indicating claim settlement agreement or claim rejection. The efficiency of car insurance claim settlement can be improved.

Description

Technical Field

[0001] The present application relates to the field of financial technology, and in particular to a motor vehicle insurance claims settlement method, a motor vehicle insurance claims settlement device, an electronic device and a storage medium. Background Art

[0002] When a vehicle is involved in a traffic accident and involves auto insurance claims, it will go through multiple links such as accident report, on-site investigation, claim document submission, vehicle assessment, claim approval, etc. The operation process of these links is cumbersome and will take a long time, resulting in low efficiency of auto insurance claims. Therefore, how to improve the efficiency of auto insurance claims has become an urgent problem to be solved. Summary of the invention

[0003] The main purpose of the embodiments of the present application is to propose a car insurance claims method, a car insurance claims device, an electronic device and a storage medium, aiming to improve the efficiency of car insurance claims.

[0004] To achieve the above-mentioned purpose, a first aspect of an embodiment of the present application proposes a method for claiming automobile insurance, the method comprising:

[0005] Obtaining the automobile insurance claim data of the target vehicle; the automobile insurance claim data includes the automobile insurance policy and driving records, the driving records include a first driving period, a first historical driving behavior, a first historical position and a first historical driving environment of the target vehicle in the first driving period, a second driving period, a second historical driving behavior, a second historical position and a second historical driving environment of the target vehicle in the second driving period; the second driving period is after the first driving period;

[0006] Extracting claim features from the motor vehicle insurance policy, the first historical driving behavior, the first historical location, and the first historical driving environment to obtain an initial claim hidden state;

[0007] Extracting claim features from the motor vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the initial claim hidden state to obtain a target claim hidden state;

[0008] The claim category of the target vehicle is determined according to the target claim hidden state; the claim category is used to indicate whether the claim is approved or rejected.

[0009] In some embodiments, extracting claim features from the auto insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the initial claim hidden state to obtain a target claim hidden state includes:

[0010] Performing feature splicing on the motor vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the initial claim hidden state to obtain an original claim splicing feature;

[0011] Acquire first impact degree data of the initial claim hidden state on the target claim hidden state according to the original claim splicing feature;

[0012] Determining a reference hidden state according to the first impact degree data and the initial claim hidden state;

[0013] Determining a first candidate hidden state according to the motor vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the original claim splicing feature;

[0014] The target claim hidden state is determined according to the first impact degree data, the reference hidden state and the first candidate hidden state.

[0015] In some embodiments, obtaining first impact degree data of the initial claim hidden state on the target claim hidden state according to the original claim splicing feature includes:

[0016] Performing linear mapping on the original claim splicing features to obtain candidate claim splicing features;

[0017] The candidate claim splicing features are activated to obtain the first impact degree data.

[0018] In some embodiments, determining the first candidate hidden state according to the auto insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the original claim splicing feature includes:

[0019] Acquire second influence degree data of the initial claim hidden state on the first candidate hidden state according to the original claim splicing feature;

[0020] Determining a baseline hidden state according to the second impact degree data and the initial claim hidden state;

[0021] The first candidate hidden state is determined according to the motor vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the reference hidden state.

[0022] In some embodiments, determining the first candidate hidden state according to the automobile insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the baseline hidden state includes:

[0023] Performing feature splicing on the motor vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the reference hidden state to obtain an initial splicing feature;

[0024] Performing linear mapping on the initial splicing features to obtain target splicing features;

[0025] The target splicing feature is activated to obtain the first candidate hidden state.

[0026] In some embodiments, determining the target claim hidden state according to the first impact degree data, the reference hidden state and the first candidate hidden state includes:

[0027] Determining a second candidate hidden state according to the first influence degree data and the first candidate hidden state;

[0028] Feature fusion is performed on the reference hidden state and the second candidate hidden state to obtain the target claim hidden state.

[0029] In some embodiments, determining the second candidate hidden state according to the first influence degree data and the first candidate hidden state includes:

[0030] Acquire third impact degree data of the first candidate hidden state on the target claim hidden state according to the first impact degree data;

[0031] The first candidate hidden state is state-weighted according to the third influence degree data to obtain the second candidate hidden state.

[0032] To achieve the above-mentioned purpose, a second aspect of an embodiment of the present application provides a vehicle insurance claims settlement device, the device comprising:

[0033] an acquisition module, configured to acquire the automobile insurance claim data of the target vehicle; the automobile insurance claim data includes the automobile insurance policy and the driving record, the driving record includes a first driving period, a first historical driving behavior, a first historical position and a first historical driving environment of the target vehicle in the first driving period, a second driving period, a second historical driving behavior, a second historical position and a second historical driving environment of the target vehicle in the second driving period; the second driving period is after the first driving period;

[0034] an initial feature extraction module, configured to extract claim features from the motor vehicle insurance policy, the first historical driving behavior, the first historical location, and the first historical driving environment to obtain an initial claim hidden state;

[0035] a target feature extraction module, configured to extract claim features from the motor vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the initial claim hidden state to obtain a target claim hidden state;

[0036] A determination module is used to determine the claim category of the target vehicle according to the target claim hidden state; the claim category is used to indicate whether the claim is approved or rejected.

[0037] To achieve the above objectives, a third aspect of an embodiment of the present application proposes an electronic device, which includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the method described in the first aspect is implemented.

[0038] To achieve the above-mentioned purpose, the fourth aspect of an embodiment of the present application proposes a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the method described in the first aspect is implemented.

[0039] The motor vehicle insurance claim settlement method, motor vehicle insurance claim settlement device, electronic device and computer-readable storage medium proposed in the embodiment of the present application obtain the motor vehicle insurance claim settlement data of the target vehicle, the motor vehicle insurance claim settlement data includes the motor vehicle insurance policy and driving record, the motor vehicle insurance policy is the basis of the motor vehicle insurance claim settlement, and the driving record is an important factor affecting the motor vehicle insurance claim settlement, so as to determine whether to settle the claim for the target vehicle based on the motor vehicle insurance policy and driving record. The claim settlement feature extraction is performed on the motor vehicle insurance policy, the first historical driving behavior, the first historical position and the first historical driving environment of the target vehicle in the first driving period, so as to extract the feature information useful for the motor vehicle insurance claim settlement in the motor vehicle insurance claim settlement data, realize effective feature extraction, and obtain the initial claim settlement hidden state. There is an intrinsic connection between the driving data such as the historical driving behavior, the historical position, and the historical driving environment in different periods of the driving record, and the driving data of multiple driving periods jointly determine the result of the motor vehicle insurance claim settlement. In order to capture the long-term dependence between the driving data of different driving periods in the driving record, the claim settlement feature extraction is performed on the motor vehicle insurance policy, the second historical driving behavior, the second historical position, the second historical driving environment and the initial claim settlement hidden state of the target vehicle in the second driving period, and the target claim settlement hidden state is obtained. The claim category of the target vehicle is determined according to the target claim hidden state. The claim category is used to indicate whether the claim is agreed or rejected, so that the claim can be automatically made for the target vehicle according to the target claim hidden state without going through a series of complicated operation processes, thereby improving the efficiency of motor vehicle insurance claims. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 is a flow chart of a method for car insurance claims settlement provided by an embodiment of the present application;

[0041] Figure 2yes Figure 1 Flow chart of step S130 in FIG.

[0042] Figure 3 yes Figure 2 Flow chart of step S220 in FIG.

[0043] Figure 4 yes Figure 2 Flow chart of step S240 in FIG.

[0044] Figure 5 yes Figure 4 Flow chart of step S430 in FIG.

[0045] Figure 6 yes Figure 2 Flow chart of step S250 in FIG.

[0046] Figure 7 yes Figure 6 Flowchart of step S610 in FIG.

[0047] Figure 8 It is a structural schematic diagram of a vehicle insurance claims settlement device provided in an embodiment of the present application;

[0048] Fig. 9 It is a schematic diagram of the hardware structure of the electronic device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0050] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0051] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0052] When a vehicle is involved in a traffic accident and involves auto insurance claims, it will go through multiple links such as accident report, on-site investigation, claim document submission, vehicle assessment, claim approval, etc. The operation process of these links is cumbersome and will take a long time, resulting in low efficiency of auto insurance claims. Therefore, how to improve the efficiency of auto insurance claims has become an urgent problem to be solved.

[0053] Based on this, the embodiments of the present application provide a motor vehicle insurance claims method, a motor vehicle insurance claims device, an electronic device and a computer-readable storage medium, aiming to improve the efficiency of motor vehicle insurance claims.

[0054] The auto insurance claims settlement method, auto insurance claims settlement device, electronic device and computer-readable storage medium provided in the embodiments of the present application are specifically illustrated through the following embodiments. First, the auto insurance claims settlement method in the embodiments of the present application is described.

[0055] The auto insurance claims method provided in the embodiment of the present application relates to the field of financial technology. The auto insurance claims method provided in the embodiment of the present application can be applied to a terminal, can be applied to a server side, or can be software running in a terminal or a server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the auto insurance claims method, etc., but is not limited to the above forms.

[0056] The present application can be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld or portable devices, tablet devices, multiprocessor systems, microprocessor-based systems, set-top boxes, programmable consumer electronics, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0057] It should be noted that in each specific implementation of the present application, when it comes to the need to perform relevant processing based on data related to user identity or characteristics such as user information, user behavior data, user historical data, and user location information, the user's permission or consent will be obtained first, and the collection, use, and processing of these data will comply with relevant laws, regulations, and standards. In addition, when the embodiment of the present application needs to obtain the user's sensitive personal information, the user's separate permission or consent will be obtained through a pop-up window or by jumping to a confirmation page. After clearly obtaining the user's separate permission or consent, the necessary user-related data for the normal operation of the embodiment of the present application will be obtained.

[0058] Figure 1 is an optional flow chart of the auto insurance claims method provided in the embodiment of the present application. Figure 1 The method may include but is not limited to steps S110 to S140.

[0059] Step S110, obtaining the automobile insurance claim data of the target vehicle; the automobile insurance claim data includes the automobile insurance policy and the driving record, the driving record includes a first driving period, a first historical driving behavior, a first historical position and a first historical driving environment of the target vehicle in the first driving period, a second driving period, a second historical driving behavior, a second historical position and a second historical driving environment of the target vehicle in the second driving period; the second driving period is after the first driving period;

[0060] Step S120, extracting claim features from the motor vehicle insurance policy, the first historical driving behavior, the first historical location, and the first historical driving environment to obtain an initial claim hidden state;

[0061] Step S130, extracting claim features from the motor vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the initial claim hidden state to obtain a target claim hidden state;

[0062] Step S140, determining the claim category of the target vehicle according to the target claim hidden state; the claim category is used to indicate whether the claim is approved or rejected.

[0063] Steps S110 to S140 shown in the embodiment of the present application capture the long-term dependencies between historical driving behaviors, historical locations, and historical driving environments in different driving periods in the driving records to achieve automated car insurance claims without going through a series of complex operational processes, thereby improving the efficiency of car insurance claims.

[0064] In step S110 of some embodiments, the target vehicle's auto insurance claim data is obtained, wherein the target vehicle is a vehicle to be claimed for auto insurance, the auto insurance claim data is the materials required when executing the auto insurance claim procedure, and the auto insurance claim data includes the auto insurance policy and driving record. The auto insurance policy includes the insurance type, insurance amount, insurance period, insured vehicle information, etc. When purchasing auto insurance, the insurance company will provide the auto insurance policy, which can be viewed and downloaded through the online service platform provided by the insurance company. The driving record is used to record the driving conditions of the target vehicle during driving, and the driving record includes the first driving period, the first historical driving behavior of the target vehicle in the first driving period, the first historical position and the first historical driving environment, the second driving period, the second historical driving behavior of the target vehicle in the second driving period, the second historical position and the second historical driving environment. The first driving period and the second driving period are the driving periods before the traffic accident of the target vehicle occurs this time, and the second driving period is after the first driving period. The first historical driving behavior, the first historical position and the first historical driving environment are respectively the driving behavior, position and driving environment of the target vehicle in the first driving period. The second historical driving behavior, the second historical position, and the second historical driving environment are respectively the driving behavior, position, and driving environment of the target vehicle in the second driving period. The vehicle is usually equipped with a driving recorder and a location navigation system, and the driving behavior and driving environment can be obtained through the driving recorder. The driving behavior includes the vehicle's speed, acceleration, braking frequency, steering wheel angle, brake pedal status, etc. The driving environment includes road conditions (such as water accumulation, bumps), weather conditions (such as fog, rain, snow), traffic conditions (such as congestion), etc. The vehicle's location can be obtained through the location navigation system.

[0065] In step S120 of some embodiments, in order to realize automated auto insurance claims, an initialization feature vector is obtained, and the auto insurance policy, the first historical driving behavior, the first historical location, the first historical driving environment and the initialization feature vector are input into the feature extraction model to extract claim features, and an initial claim hidden state is obtained. The feature extraction model may be a gated recurrent unit (GRU), and the initial claim hidden state represents the feature information of the target vehicle in the first driving period learned by the feature extraction model during the feature extraction process. A zero vector may be used as the initialization feature vector, or multiple random values ​​may be extracted from a preset data distribution, and an initialization feature vector may be constructed based on the multiple random values. The preset data distribution may be a Gaussian distribution, a uniform distribution, etc., and multiple random values ​​may be extracted from a Gaussian distribution with a mean of 0 and a standard deviation of a given value, the given value may be 0.01 or 0.1, or multiple random values ​​may be extracted from a uniform distribution with a data range of [-a, a], where a is a small positive number, such as 0.1.

[0066] See also Figure 2In some embodiments, step S130 may include but is not limited to steps S210 to S250:

[0067] Step S210, performing feature splicing on the motor vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the initial claim hidden state to obtain original claim splicing features;

[0068] Step S220, obtaining first impact degree data of the initial claim hidden state on the target claim hidden state according to the original claim splicing feature;

[0069] Step S230, determining a reference hidden state according to the first impact degree data and the initial claim hidden state;

[0070] Step S240, determining a first candidate hidden state according to the automobile insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the original claim splicing feature;

[0071] Step S250, determining a target claim hidden state according to the first impact degree data, the reference hidden state and the first candidate hidden state.

[0072] In step S210 of some embodiments, driving data such as driving behavior, position, and driving environment in different driving periods in the driving record form a time series. The driving data changes dynamically with different driving periods, making the time series time-dependent. There is an intrinsic correlation between driving data in driving periods with long intervals, making the time series long-term dependent. In order to capture the time dependency and long-term dependency in the time series, the claim feature extraction is performed on the motor vehicle insurance policy, the second historical driving behavior, the second historical position, the second historical driving environment, and the initial claim hidden state of the target vehicle in the second driving period. Specifically, feature splicing is performed on the motor vehicle insurance policy, the second historical driving behavior, the second historical position, the second historical driving environment, and the initial claim hidden state, and different feature information is integrated to provide a more comprehensive and richer feature representation, and the original claim splicing feature is obtained.

[0073] During the auto insurance claims process, the initial claim hidden state only focuses on the driving characteristics of the target vehicle in the first driving period, ignoring the driving characteristics of the target vehicle in the second driving period and the dependencies between different driving periods, so that the initial claim hidden state only contains local features, while the original claim splicing features take into account the dynamic data characteristics of the entire vehicle driving process, providing a more comprehensive feature representation, thereby improving the accuracy of auto insurance claims prediction.

[0074] In step S220 of some embodiments, in order to solve the problems of gradient vanishing and difficulty in modeling long-term dependencies in traditional recurrent neural networks (RNN), long short-term memory networks (LSTM) are usually used to extract long-term dependencies in time series. The long short-term memory network introduces an input gate, a forget gate, an output gate, and an additional unit state, which makes the network have a large number of parameters and requires more computing resources and time for claim feature extraction. It takes more than 20 seconds to predict the claim category of a vehicle, resulting in low efficiency in auto insurance claims and reducing the claims experience of car owners. In addition, when dealing with long-term dependency problems, the long short-term memory network tends to over-focus on the last input data, reducing the accuracy of auto insurance claims. For example, the driving record includes the first driving data of the target vehicle in the first driving period, the second driving data in the second driving period, and the third driving data in the third driving period. The second driving period is after the first driving period, and the third driving period is after the second driving period. The hidden states of claims extracted based on the first driving data and the second driving data both indicate that the claim category is rejected, and the hidden state of claims extracted based on the third driving data indicates that the claim category is agreed. LSTM will ignore the first two rejected claims and only focus on the last claim, resulting in low accuracy of auto insurance claims.

[0075] In order to improve the efficiency and accuracy of auto insurance claims, the embodiment of the present application optimizes the time series through a gating mechanism to control the transmission of information, so as to more effectively capture long-term and short-term dependencies and predict more diverse time dynamic features. The gating mechanism includes an update gate and a reset gate, which retain important historical hidden state information. For example, in three claims, more attention will be paid to the first two rejections. The update gate is used to capture long-term dependencies in the sequence, and the reset gate is used to capture short-term dependencies in the sequence. The gating mechanism merges the input gate and the forget gate in the LSTM model into an update gate, which simplifies the structure of the LSTM model, making the convergence speed of the gating mechanism faster than the LSTM model. Under the same data and training time, the gating mechanism reaches the optimal performance faster to achieve real-time auto insurance claims.

[0076] The update gate determines how much proportion of the initial claim hidden state is retained in the target claim hidden state. The original claim concatenation feature is input into the update gate to calculate the first impact degree data of the initial claim hidden state on the target claim hidden state. The first impact degree data is the output parameter of the update gate, which is used to indicate the proportion of the initial claim hidden state in the target claim hidden state. The first impact degree data is a value between [0,1].

[0077] In step S230 of some embodiments, the first impact degree data and the initial claim hidden state are multiplied to obtain a reference hidden state.

[0078] In step S240 of some embodiments, in order to capture complex patterns and dependencies in the driving record, candidate hidden states are calculated based on the auto insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the original claim splicing features to obtain a first candidate hidden state. The first candidate hidden state is a potential new state adjusted by the reset gate, not a real hidden state, which is used to generate a real hidden state, i.e., a target claim hidden state, obtained by combining the initial claim hidden state and the current input data.

[0079] In step S250 of some embodiments, the embodiments of the present application do not introduce additional unit states like the LSTM model, but directly construct a linear dependency relationship based on the first impact degree data, the reference hidden state and the first candidate hidden state to obtain the target claim hidden state, thereby avoiding additional calculations and improving the efficiency of motor insurance claims.

[0080] Through the above steps S210 to S250, it is possible to capture the long-term and short-term dependencies in the time series, determine the time-varying interactions between driving behavior, location, and driving environment in the auto insurance claim data, and the impact of these factors on the auto insurance claim, and obtain the target claim hidden state.

[0081] See also Figure 3 In some embodiments, step S220 may include but is not limited to steps S310 to S320:

[0082] Step S310, linear mapping is performed on the original claim splicing features to obtain candidate claim splicing features;

[0083] Step S320, activating the candidate claim splicing features to obtain first impact degree data.

[0084] In step S310 of some embodiments, the update gate includes a weight matrix, the weight matrix includes weights of the auto insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the initial claim hidden state. The weight matrix is ​​multiplied by the original claim splicing feature to obtain the candidate claim splicing feature.

[0085] In step S320 of some embodiments, the candidate claim splicing features are activated by a sigmoid activation function to obtain first impact degree data. The calculation formula of the first impact degree data is as follows:

[0086] z t =σ(W z ·[h t-1,x t ]),

[0087] Among them, z t represents the first influence degree data; σ represents the sigmoid activation function; W z represents the weight matrix; represents the matrix multiplication operation; h t-1 Indicates the initial claim hidden state; x t represents the current input data, including the car insurance policy, the second historical driving behavior, the second historical location and the second historical driving environment, t-1 represents the first driving period, and t represents the second driving period.

[0088] The update gate determines how much of the target claim hidden state comes from the initial claim hidden state (old information) and how much comes from the current input data x t (new information) to weigh between new information and old information, so as to decide what information needs to be retained and what needs to be updated. When the first impact degree data is close to 1, it tends to retain and not update the initial claim hidden state. When the first impact degree data is close to 0, it means that the initial claim hidden state is not retained, and the hidden state is updated based on the current input data. For example, the first impact degree data is 0.82, which means that the initial claim hidden state is important information, and the initial claim hidden state will not be replaced according to the current input auto insurance claim data, ensuring the accuracy of auto insurance claims.

[0089] Through the above steps S310 to S320, the first impact degree data can be determined to control the transfer of the initial claim hidden state based on the first impact degree data, thereby completing the modeling of the long-term dependency relationship.

[0090] See also Figure 4 In some embodiments, step S240 may include but is not limited to steps S410 to S430:

[0091] Step S410, obtaining second impact degree data of the initial claim hidden state on the first candidate hidden state according to the original claim splicing feature;

[0092] Step S420, determining a baseline hidden state according to the second impact degree data and the initial claim hidden state;

[0093] Step S430, determining a first candidate hidden state according to the automobile insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment and the reference hidden state.

[0094] In step S410 of some embodiments, the reset gate controls the influence of the initial claim hidden state on the first candidate hidden state, inputs the original claim splicing feature into the reset gate, and calculates the second influence degree data of the initial claim hidden state on the first candidate hidden state, the second influence degree data is the output parameter of the reset gate, and is used to indicate the proportion of the initial claim hidden state in the first candidate hidden state. The second influence degree data is a value between [0,1].

[0095] Specifically, the weight matrix of the reset gate is obtained, the weight matrix is ​​multiplied by the original claim splicing feature, and the multiplied feature is activated by the sigmoid activation function to obtain the second impact degree data. The weight matrix includes the weights of each feature element in the original claim splicing feature, such as the weights of the auto insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the initial claim hidden state. The calculation formula for the second impact degree data is as follows:

[0096] r t =σ(W r ·[h t-1 ,x t ]),

[0097] Among them, r t represents the second influence degree data; σ represents the sigmoid activation function; W r represents the weight matrix of the reset gate; represents the matrix multiplication operation; h t-1 Indicates the initial claim hidden state; x t represents the current input data, including the car insurance policy, the second historical driving behavior, the second historical location and the second historical driving environment, t-1 represents the first driving period, and t represents the second driving period.

[0098] In step S420 of some embodiments, the second impact degree data is multiplied by the initial claim hidden state to obtain a reference hidden state. t , the initial claim hidden state is represented by h t-1 , then the base hidden state is represented by r t ·h t-1 ,· represents the matrix multiplication operation.

[0099] When the second impact degree data output by the reset gate is close to 0, the initial claim hidden state has little impact on the first candidate hidden state, tends to ignore previous historical information, and mainly relies on the current input data. When the second impact degree data output by the reset gate is close to 1, the initial claim hidden state has a greater impact on the first candidate hidden state, retaining more historical information. For example, if the second impact degree data is 0.79, choose to retain more historical information.

[0100] In step S430 of some embodiments, a nonlinear transformation is performed on the vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the reference hidden state to obtain a first candidate hidden state.

[0101] In the above steps S410 to S430, by controlling the influence of the initial claim hidden state on the first candidate hidden state through the second influence degree data, it can be determined whether to retain the initial claim hidden state, thereby more effectively handling the long-term dependency problem and determining the intrinsic correlation between driving data in different driving periods, so as to accurately assess the accident risk of the driver under different driving conditions and obtain more accurate motor vehicle insurance claim prediction results.

[0102] See also Figure 5 In some embodiments, step S430 may include but is not limited to steps S510 to S530:

[0103] Step S510, performing feature splicing on the vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the reference hidden state to obtain an initial splicing feature;

[0104] Step S520, performing linear mapping on the initial splicing features to obtain target splicing features;

[0105] Step S530, activating the target splicing feature to obtain a first candidate hidden state.

[0106] In step S510 of some embodiments, the car insurance policy, the second historical driving behavior, the second historical location, and the second historical driving environment are current input data, represented by x t , the base hidden state is denoted as r t ·h t-1 , concatenate the baseline hidden state and the current input data to obtain the initial concatenated feature [r t ·h t-1 ,x t ].

[0107] In step S520 of some embodiments, a weight matrix for calculating the first candidate hidden state is obtained, and the weight matrix is ​​multiplied by the initial splicing feature to obtain the target splicing feature.

[0108] In step S530 of some embodiments, the target splicing feature is activated by a hyperbolic tangent function tanh to obtain a first candidate hidden state. The calculation formula of the first candidate hidden state is as follows:

[0109] h′ t =tanh(W h ·[r t ·h t-1 ,xt ]),

[0110] Among them, h′ t represents the first candidate hidden state; tanh represents the hyperbolic tangent function; W h represents the weight matrix; r t Indicates the second impact level data; h t-1 Indicates the initial claim hidden state; x t Represents the current input data; Represents the matrix multiplication operation.

[0111] Through the above steps S510 to S530, the hidden state can be continuously updated over time to obtain a potential new state, namely the first candidate hidden state, to reflect the driver's latest driving risk status. This real-time update mechanism can adapt to dynamically changing driving records, making the prediction results of auto insurance claims more accurate.

[0112] See also Figure 6 In some embodiments, step S250 may include but is not limited to steps S610 to S620:

[0113] Step S610, determining a second candidate hidden state according to the first influence degree data and the first candidate hidden state;

[0114] Step S620, performing feature fusion on the reference hidden state and the second candidate hidden state to obtain the target claim hidden state.

[0115] In step S610 of some embodiments, a linear transformation is performed on the first candidate hidden state according to the first influence degree data to obtain a second candidate hidden state.

[0116] In step S620 of some embodiments, the reference hidden state and the second candidate hidden state are feature-aligned and added to obtain a target claim hidden state.

[0117] The above steps S610 to S620, by combining the reference hidden state and the second candidate hidden state, can retain important historical information when updating the initial claim hidden state, and introduce new information to accurately reflect the driving risk conditions at different driving times.

[0118] See also Figure 7 In some embodiments, step S610 may include but is not limited to steps S710 to S720:

[0119] Step S710, obtaining third impact degree data of the first candidate hidden state on the target claim hidden state according to the first impact degree data;

[0120] Step S720: weight the first candidate hidden state according to the third influence degree data to obtain a second candidate hidden state.

[0121] In step S710 of some embodiments, the given value is 1, and the first impact degree data is subtracted from the given value to obtain the third impact degree data. The third impact degree data is used to indicate the proportion of the first candidate hidden state in the target claim hidden state. The third impact degree data is a value between [0, 1], and its specific value is related to the first impact degree data, and the sum of the first impact degree data and the third impact degree data is 1.

[0122] In step S720 of some embodiments, the third influence degree data is multiplied by the first candidate hidden state to obtain the second candidate hidden state.

[0123] Through the above steps S710 to S720, new status information can be obtained, and the initial hidden claim status can be updated using the new status information, so that the hidden claim status changes over time, so as to accurately assess the driving risk situation under different driving periods, thereby ensuring the accuracy of the motor vehicle insurance claims process.

[0124] In step S140 of some embodiments, the target claim hidden state is identified by a classification model to obtain the claim category of the target vehicle. The claim category is used to indicate whether the claim is approved or rejected. The classification model can be a multi-layer perceptron, a decision tree, a support vector machine, etc. If the claim category is approved, the target claim hidden state can also be input into a regression model to predict the claim amount of the target vehicle. The regression model can be a linear regression model, a random forest regression model, etc.

[0125] The motor vehicle insurance claim settlement method of the embodiment of the present application includes: obtaining the motor vehicle insurance claim settlement data of the target vehicle, the motor vehicle insurance claim settlement data including the motor vehicle insurance policy, the driving behavior, position and driving environment of the target vehicle in multiple driving periods, and the multiple driving periods are arranged in order from the first to the last. According to the order of the driving periods, the initialization feature vector is obtained, and the claim settlement hidden state of the first driving period is determined according to the initialization feature vector, the motor vehicle insurance policy, the driving behavior, position and driving environment of the target vehicle in the first driving period. Taking the second driving period as the starting point, the motor vehicle insurance policy, the driving behavior, position and driving environment of the target vehicle in the current driving period are used as the current input data, and the claim settlement hidden state of the previous driving period and the current input data are spliced ​​to obtain the original claim settlement splicing feature. The original claim settlement splicing feature is feature mapped by the update gate to obtain the first influence degree data of the claim settlement hidden state of the previous driving period on the claim settlement hidden state of the current driving period. The original claim settlement splicing feature is feature mapped by the reset gate to obtain the second influence degree data of the claim settlement hidden state of the previous driving period on the candidate hidden state of the current driving period. Multiply the second impact degree data by the claim hidden state of the previous driving period to obtain a baseline hidden state. Perform feature splicing on the baseline hidden state and the current input data to obtain an initial splicing feature. Perform linear mapping on the initial splicing feature to obtain a target splicing feature. Activate the target splicing feature through the hyperbolic tangent function to obtain a candidate hidden state for the current driving period. Multiply the first impact degree data by the claim hidden state of the previous driving period, multiply the preset third impact degree data by the candidate hidden state of the current driving period, and add the two multiplication results to obtain the claim hidden state of the current driving period. The sum of the first impact degree data and the third impact degree data is 1. Repeat the above steps until the claim hidden state of the last driving period is obtained. Determine the claim category of the target vehicle based on the claim hidden state of the last driving period, and the claim category is used to indicate whether the claim is agreed or rejected.

[0126] The embodiment of the present application simplifies the structure of the LSTM model through the update gate and the reset gate, thereby simplifying the process of updating the hidden state of the claim, avoiding the consumption of a large amount of computing resources and costs, and improving the efficiency of motor vehicle insurance claims. In addition, the update gate and the reset gate process the long-term and short-term dependencies of time series data by controlling the transmission of information, thereby enhancing the flexibility of updating the hidden state of the claim. The driving behavior, position, and driving environment will be continuously updated with the driving period, and there is an intrinsic correlation between the driving behavior, position, and driving environment in different driving periods. The hidden state of the claim of the current driving period is determined by combining the current input data with the hidden state of the claim of the previous driving period to establish the time dependency relationship between the hidden states of the claim of different driving periods, thereby more accurately predicting the accident risk under different driving conditions, and ensuring the accuracy of the prediction results of motor vehicle insurance claims.

[0127] See also Figure 8 The embodiment of the present application also provides a vehicle insurance claim settlement device, which can implement the above vehicle insurance claim settlement method, and the device includes:

[0128] The acquisition module 810 is used to acquire the automobile insurance claim data of the target vehicle; the automobile insurance claim data includes the automobile insurance policy and the driving record, and the driving record includes a first driving period, a first historical driving behavior of the target vehicle in the first driving period, a first historical position and a first historical driving environment, a second driving period, a second historical driving behavior of the target vehicle in the second driving period, a second historical position and a second historical driving environment; the second driving period is after the first driving period;

[0129] An initial feature extraction module 820 is used to extract claim features from the motor vehicle insurance policy, the first historical driving behavior, the first historical location, and the first historical driving environment to obtain an initial claim hidden state;

[0130] A target feature extraction module 830 is used to extract claim features from the motor vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the initial claim hidden state to obtain a target claim hidden state;

[0131] The determination module 840 is used to determine the claim category of the target vehicle according to the target claim hidden state; the claim category is used to indicate whether the claim is approved or rejected.

[0132] The specific implementation of the vehicle insurance claims settlement device is basically the same as the specific implementation of the above-mentioned vehicle insurance claims settlement method, and will not be repeated here.

[0133] The embodiment of the present application also provides an electronic device, the electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the above-mentioned auto insurance claim settlement method when executing the computer program. The electronic device can be any smart terminal including a tablet computer, a car computer, etc.

[0134] See also Fig. 9 , Fig. 9 The hardware structure of an electronic device of another embodiment is illustrated, and the electronic device includes:

[0135] The processor 910 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (Application Specific Integrated Circuit, ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0136] The memory 920 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 920 can store an operating system and other application programs. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 920, and the processor 910 calls and executes the auto insurance claim settlement method of the embodiment of this application;

[0137] Input / output interface 930, used to implement information input and output;

[0138] Communication interface 940, used to realize communication interaction between the device and other devices, which can be realized through wired mode (such as USB, network cable, etc.) or wireless mode (such as mobile network, WIFI, Bluetooth, etc.);

[0139] bus 950 , which transmits information between the various components of the device (e.g., processor 910 , memory 920 , input / output interface 930 , and communication interface 940 );

[0140] The processor 910 , the memory 920 , the input / output interface 930 , and the communication interface 940 are connected to each other in communication within the device via a bus 950 .

[0141] An embodiment of the present application also provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the above-mentioned car insurance claim settlement method is implemented.

[0142] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0143] The car insurance claims method, car insurance claims device, electronic device and computer storage medium provided in the embodiments of the present application realize automated car insurance claims by capturing the long-term dependencies between historical driving behaviors, historical locations and historical driving environments in different driving periods in driving records, without going through a series of complex operational processes, thereby improving the efficiency of car insurance claims.

[0144] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0145] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0146] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0147] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0148] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0149] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0150] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0151] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0152] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0153] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0154] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but the scope of the rights of the present invention is not limited thereto. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present invention should be within the scope of the rights of the present invention.

Claims

1. A method for settling claims for automobile insurance, characterized in that: The method comprises: Obtaining the automobile insurance claim data of the target vehicle; the automobile insurance claim data includes the automobile insurance policy and driving records, the driving records include a first driving period, a first historical driving behavior, a first historical position and a first historical driving environment of the target vehicle in the first driving period, a second driving period, a second historical driving behavior, a second historical position and a second historical driving environment of the target vehicle in the second driving period; the second driving period is after the first driving period; Extracting claim features from the motor vehicle insurance policy, the first historical driving behavior, the first historical location, and the first historical driving environment to obtain an initial claim hidden state; Extracting claim features from the motor vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the initial claim hidden state to obtain a target claim hidden state; The claim category of the target vehicle is determined according to the target claim hidden state; the claim category is used to indicate whether the claim is approved or rejected.

2. The method according to claim 1, characterized in that The extracting the claim feature of the motor vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the initial claim hidden state to obtain a target claim hidden state includes: Performing feature splicing on the motor vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the initial claim hidden state to obtain an original claim splicing feature; Acquire first impact degree data of the initial claim hidden state on the target claim hidden state according to the original claim splicing feature; Determining a reference hidden state according to the first impact degree data and the initial claim hidden state; Determining a first candidate hidden state according to the motor vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the original claim splicing feature; The target claim hidden state is determined according to the first impact degree data, the reference hidden state and the first candidate hidden state.

3. The method according to claim 2, characterized in that The obtaining, according to the original claim splicing feature, first impact degree data of the initial claim hidden state on the target claim hidden state includes: Performing linear mapping on the original claim splicing features to obtain candidate claim splicing features; The candidate claim splicing feature is activated to obtain the first impact degree data.

4. The method according to claim 2, characterized in that: The determining of the first candidate hidden state according to the motor vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the original claim splicing feature includes: Acquire second influence degree data of the initial claim hidden state on the first candidate hidden state according to the original claim splicing feature; Determining a baseline hidden state according to the second impact degree data and the initial claim hidden state; The first candidate hidden state is determined according to the motor vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the reference hidden state.

5. The method according to claim 4, characterized in that The determining the first candidate hidden state according to the automobile insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the reference hidden state includes: Performing feature splicing on the motor vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the reference hidden state to obtain an initial splicing feature; Performing linear mapping on the initial splicing features to obtain target splicing features; The target splicing feature is activated to obtain the first candidate hidden state.

6. The method according to claim 2, characterized in that The step of determining the target claim hidden state according to the first impact degree data, the reference hidden state and the first candidate hidden state includes: Determining a second candidate hidden state according to the first influence degree data and the first candidate hidden state; Feature fusion is performed on the reference hidden state and the second candidate hidden state to obtain the target claim hidden state.

7. The method according to claim 6, characterized in that The determining a second candidate hidden state according to the first influence degree data and the first candidate hidden state includes: Acquire third impact degree data of the first candidate hidden state on the target claim hidden state according to the first impact degree data; The first candidate hidden state is state-weighted according to the third influence degree data to obtain the second candidate hidden state.

8. A car insurance claims device, characterized in that: The device comprises: an acquisition module, configured to acquire the automobile insurance claim data of the target vehicle; the automobile insurance claim data includes the automobile insurance policy and the driving record, the driving record includes a first driving period, a first historical driving behavior, a first historical position and a first historical driving environment of the target vehicle in the first driving period, a second driving period, a second historical driving behavior, a second historical position and a second historical driving environment of the target vehicle in the second driving period; the second driving period is after the first driving period; an initial feature extraction module, configured to extract claim features from the motor vehicle insurance policy, the first historical driving behavior, the first historical location, and the first historical driving environment to obtain an initial claim hidden state; a target feature extraction module, configured to extract claim features from the motor vehicle insurance policy, the second historical driving behavior, the second historical location, the second historical driving environment, and the initial claim hidden state to obtain a target claim hidden state; A determination module is used to determine the claim category of the target vehicle according to the target claim hidden state; the claim category is used to indicate whether the claim is approved or rejected.

9. An electronic device, characterized in that: The electronic device comprises a memory and a processor, the memory stores a computer program, and the processor implements the method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the method according to any one of claims 1 to 7 is implemented.

Images