WO2020024406A1

WO2020024406A1 - Electronic device, method for scheduling car insurance investigation tasks based on traffic factor, and storage medium

Info

Publication number: WO2020024406A1
Application number: PCT/CN2018/107708
Authority: WO
Inventors: 朱菊花
Original assignee: 平安科技（深圳）有限公司
Priority date: 2018-08-01
Filing date: 2018-09-26
Publication date: 2020-02-06
Also published as: CN109300046A

Abstract

An electronic device, a method for scheduling car insurance investigation tasks based on a traffic factor, and a storage medium. The method comprises: after receiving a claim request associated with a car insurance case, acquiring information of the car insurance case (S100); determining an investigation task for the case and investigators matching the investigation task, and respectively acquiring second geographical locations at which the respective matched investigators are currently located (200); respectively determining travel routes between a first geographical location and the second geographical locations, and acquiring traffic factors of the respective determined travel routes; inputting the acquired traffic factors of the respective travel routes to a gradient boosting decision tree model to decide a travel route with the least travel time from the respective determined travel routes (S300); and determining the investigator and assigning the investigation task for the case thereto (S400). The invention can eliminate unexpected factors that prevent investigators from arriving accident sites on time according to traffic factors, thereby ensuring that cases are timely and accurately dealt with, and accordingly increasing service satisfaction of clients.

Description

Electronic device, vehicle insurance survey and dispatch method based on road condition factor, and storage medium

This application claims the priority of a Chinese patent application filed on August 1, 2018 with the Chinese Patent Office, application number 201810862694.8, and the invention name is "electronic device, road condition factor-based vehicle insurance survey and dispatch method and storage medium", and its entire contents Incorporated by reference in this application.

Technical field

The present application relates to the field of automobile insurance concessions, and in particular, to an electronic device, a method for automobile insurance survey and dispatch based on road condition factors, and a storage medium.

Background technique

With the popularization of insurance business in various industries, vehicle insurance has penetrated into almost every vehicle user. However, in the prior art, when a vehicle accident requires the assignment of investigation tasks, often only the case information of the outbreak case is considered, such as the case information of the outbreak case, the type of the case, the degree of damage and other case information, and the inspector to the outbreak is ignored. The road condition factor of the location may cause the investigator to arrive on time due to environmental factors such as road conditions, which delays the timely processing of the case and affects the customer's satisfaction with the service.

Summary of the invention

In view of this, the present application proposes an electronic device, the electronic device includes a memory, and a processor connected to the memory, the processor is configured to execute a road insurance factor-based vehicle insurance survey and scheduling program stored on the memory, When the vehicle risk survey and dispatch program based on the road condition factor is executed by the processor, the following steps are implemented:

A10. After receiving the auto insurance report request, obtain auto insurance case information, where the case information includes the first geographic location of the case and the type of the case;

A20: Query the mapping table between the pre-stored case types, survey tasks, and surveyors based on the case types to determine the survey tasks of the case and the surveyors that match the survey tasks, and obtain the matching surveys. The second geographic location where the employee is currently located;

A30. Determine travel routes between the first geographic location and the second geographic location, respectively, and obtain road condition factors on the determined travel routes.

A40. Input the obtained road condition factors on each travel route into a gradient boosting decision tree model to obtain prediction result data output from the gradient boosting decision tree model, where the prediction result data is from the determined travel routes, The shortest travel route to be decided;

A50. According to the second geographical location corresponding to the shortest required travel route determined by the decision, it is determined that the investigation task of the case is assigned to the surveyor in the second geographical location.

In addition, in order to achieve the above object, the present application also proposes a vehicle insurance survey and dispatch method based on road condition factors, which is characterized in that the method includes the following steps:

S10. After receiving the auto insurance report request, obtain auto insurance case information, where the case information includes the first geographic location of the outbreak case and the case type;

S20: Query the mapping table between the pre-stored case type, the survey task, and the surveyor based on the case type to determine the survey task of the case and each surveyor matching the survey task, and obtain each matching survey The second geographic location where the employee is currently located;

S30. Determine travel routes between the first geographic location and the second geographic location, respectively, and obtain road condition factors on the determined travel routes.

S40. Input the obtained road condition factors on each travel route into a gradient promotion decision tree model to obtain prediction result data output by the gradient promotion decision tree model, where the prediction result data is from the determined travel routes, The shortest travel route to be decided;

S50. According to the second geographical location corresponding to the shortest required travel route determined by the decision, it is determined that the investigation task of the case is assigned to the surveyor in the second geographical location.

In addition, in order to achieve the above object, the present application also proposes a computer-readable storage medium storing a vehicle insurance survey and dispatching program based on road condition factors, and the vehicle insurance survey and dispatching program based on road condition factors may be at least A processor executes to cause the at least one processor to perform the following steps:

After receiving the auto insurance report request, obtain auto insurance case information, where the case information includes the first geographic location of the outbreak case and the case type;

Based on the case type, query the pre-stored mapping table between the case type, the survey task, and the surveyor to determine the survey task for the case and each surveyor that matches the survey task, and obtain the matching surveyor current Second geographical location

Determining travel routes between the first geographic location and the second geographic location, respectively, and acquiring traffic condition factors on the determined travel routes;

The obtained traffic condition factors on each travel route are input to a gradient boosting decision tree model to obtain prediction result data output by the gradient boosting decision tree model, where the prediction result data is determined from the determined travel routes The shortest travel route required;

According to the second geographic location corresponding to the shortest required travel route determined by the decision, it is determined that the investigation task of the case is assigned to the surveyor in the second geographic location.

The electronic device, the road insurance factor-based vehicle insurance survey and dispatch method, and the storage medium provided in this application first obtain vehicle insurance case information after receiving a vehicle insurance report request, and the case information includes the first geographic location of the outbreak case and the case type; Secondly, based on the case type, query the mapping relationship table between the pre-stored case type and the survey task and the surveyor to determine the survey task of the case and each surveyor matching the survey task, and obtain each matching surveyor. The second geographic location where the current location is currently located; then each travel route between the first geographic location and the second geographic location is determined separately, and the determined traffic condition factors on each of the travel routes are obtained; The traffic condition factor is input to the gradient boosting decision tree model to obtain the prediction result data output by the gradient boosting decision tree model, where the prediction result data is the shortest travel time determined from the determined travel routes. Route; the shortest travel route pair based on the decision The second location, survey tasks to determine the case in the survey distributed to members in the second geographical location. It can exclude accidental factors that prevent investigators from arriving on time according to road conditions factors, improve the timely accuracy of case processing, and improve customer satisfaction with services.

BRIEF DESCRIPTION OF THE DRAWINGS

FIG. 1 is a schematic diagram of an optional hardware architecture of an electronic device proposed by the present application; FIG.

2 is a schematic diagram of a program module of a vehicle insurance survey and dispatching schedule based on a road condition factor in an embodiment of an electronic device of the present application;

FIG. 3 is an implementation flowchart of a preferred embodiment of a vehicle insurance survey and dispatch method based on road condition factors of the present application.

The implementation, functional features and advantages of the purpose of this application will be further described with reference to the embodiments and the drawings.

detailed description

In order to make the purpose, technical solution, and advantages of the present application clearer, the present application is further described in detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the application, and are not used to limit the application. Based on the embodiments in the present application, all other embodiments obtained by a person of ordinary skill in the art without creative efforts shall fall within the protection scope of the present application.

It should be noted that the descriptions related to "first" and "second" in this application are only for descriptive purposes, and cannot be understood as indicating or implying their relative importance or implicitly indicating the number of technical features indicated. . Therefore, the features defined as "first" and "second" may explicitly or implicitly include at least one of the features. In addition, the technical solutions between the various embodiments can be combined with each other, but must be based on those that can be realized by a person of ordinary skill in the art. When the combination of technical solutions conflicts or cannot be achieved, it should be considered that such a combination of technical solutions does not exist. Is not within the scope of protection claimed in this application.

Refer to FIG. 1, which is a schematic diagram of an optional hardware architecture of an electronic device according to the present application. In this embodiment, the electronic device 10 may include, but is not limited to, a memory 11, a processor 12, and a network interface 13 which may communicate with each other through a communication bus 14. It should be noted that FIG. 1 only shows the electronic device 10 having components 11-14, but it should be understood that it is not required to implement all the illustrated components, and more or fewer components may be implemented instead.

The memory 11 includes at least one type of computer-readable storage medium. The computer-readable storage medium includes a flash memory, a hard disk, a multimedia card, a card-type memory (for example, SD or DX memory, etc.), a random access memory (RAM), and a static memory. Random access memory (SRAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), programmable read-only memory (PROM), magnetic memory, magnetic disks, optical disks, etc. In some embodiments, the memory 11 may be an internal storage unit of the electronic device 10, such as a hard disk or a memory of the electronic device 10. In other embodiments, the memory 11 may also be an outsourced storage device of the electronic device 10, such as a plug-in hard disk, a smart memory card (SMC), and a secure digital (SD) device. ) Cards, flash cards, etc. Of course, the memory 11 may also include both the internal storage unit of the electronic device 10 and its outsourced storage device. In this embodiment, the memory 11 is generally used to store an operating system and various application software installed on the electronic device 10, such as a car insurance survey and dispatching program based on a road condition factor. In addition, the memory 11 can also be used to temporarily store various types of data that have been output or will be output.

The processor 12 may be a central processing unit (CPU), a controller, a microcontroller, a microprocessor, or other data processing chip in some embodiments. The processor 12 is generally used to control the overall operation of the electronic device 10. In this embodiment, the processor 12 is configured to run program code or process data stored in the memory 11, for example, a vehicle insurance survey scheduler based on a road condition factor and the like.

The network interface 13 may include a wireless network interface or a wired network interface. The network interface 13 is generally used to establish a communication connection between the electronic device 10 and other electronic devices.

The communication bus 14 is used to implement a communication connection between the components 11-13.

FIG. 1 only shows the electronic device 10 with components 11-14 and a road insurance factor-based vehicle insurance survey scheduler, but it should be understood that it is not required to implement all of the components shown, and more or less can be implemented instead. s component.

Optionally, the electronic device 10 may further include a user interface (not shown in FIG. 1). The user interface may include a display, an input unit such as a keyboard, and the user interface may further include a standard wired interface, a wireless interface, and the like.

Optionally, in some embodiments, the display may be an LED display, a liquid crystal display, a touch-type liquid crystal display, an OLED touch device, or the like. Further, the display may also be referred to as a display screen or a display unit, for displaying information processed in the electronic device 10 and for displaying a visualized user interface.

Optionally, in some embodiments, the electronic device 10 may further include an audio unit (the audio unit is not shown in FIG. 1), and the audio unit may be in the electronic device 10 in a call signal receiving mode, a call mode, a recording mode, and voice recognition. In the mode, the broadcast receiving mode, and the like, the received or stored audio data is converted into an audio signal. Further, the electronic device 10 may further include an audio output unit, and the audio output unit outputs the audio signal converted by the audio unit, and The audio output unit may also provide audio output (such as call signal reception sound, message reception sound, etc.) related to a specific function performed by the electronic device 10, and the audio output unit may include a speaker, a buzzer, and the like.

Optionally, in some embodiments, the electronic device 10 may further include an alarm unit (not shown in the figure), and the alarm unit may provide an output to notify the electronic device 10 of the occurrence of the event. Typical events may include call reception, message reception, key signal input, touch input, and so on. In addition to audio or video output, the alarm unit can provide output in different ways to notify the occurrence of an event. For example, the alarm unit may provide an output in the form of a vibration, and when a call, message, or some other may cause the electronic device 10 to enter the communication mode, the alarm unit may provide a tactile output (ie, vibration) to notify the user.

In an embodiment, when the vehicle risk survey and dispatching program based on the road condition factor stored in the memory 11 is executed by the processor 12, the following operations are implemented:

A. After receiving the auto insurance report request, obtain auto insurance case information, where the case information includes the first geographic location of the case and the type of the case;

Specifically, if the agent receives the call of the car insurance case reporter, the car insurance case reporter informs the agent of the relevant case information and enters the relevant case information, that is, receives the car insurance case information. The corresponding auto insurance case information can also be extracted according to the language information of the car insurance case reporter. Preferably, the case information further includes an affiliated insurance company, the name of the damaged component, the degree of damage, the location of the damage, and the like.

B. Query the mapping table between the pre-stored case types, survey tasks, and surveyors based on the case types to determine the survey tasks of the case and the surveyors that match the survey tasks, and obtain the matching surveys. The second geographic location where the employee is currently located;

C. Determine travel routes between the first geographic location and the second geographic location, respectively, and obtain road condition factors on the determined travel routes;

Specifically, a request for obtaining the determined road condition factors on each travel route may be sent to the traffic database of the traffic supervision department to obtain the determined road condition factors on each travel route. The road condition factor may be recorded in a picture or a video file, and the road condition factor may include road subgrade, pavement, and surrounding buildings on the travel route, for example, the type of road subgrade and the degree of potholes on the road. , The height of buildings around the road, water on the pavement, etc., can also include road traffic conditions on the travel route, for example, the intensive situation of vehicles driving on the road, the dense flow of people, the degree of road traffic congestion, etc., and can also include Video and sound in the travel route. The road condition factor obtained in this embodiment is stored in a traffic database of a traffic supervision department, and picture information and / or image information obtained through a camera with a high-definition resolution is captured.

D. Input the obtained road condition factors on each travel route into a gradient boosting decision tree model to obtain prediction result data output by the gradient boosting decision tree model, where the prediction result data is from the determined travel routes, The shortest travel route to be decided;

Specifically, the gradient boosting decision tree model is referred to as GBDT (Gradient Boosting Decision Tree), also called MART (Multiple Additive Regression Tree), GBT (Gradient Boosting Tree), GBT (Gradient Tree Boosting), or GBRT (Gradient Boosting Regression Tree). In this embodiment, they are collectively referred to as GBDT. The GBDT is an iterative decision tree algorithm. The algorithm can be composed of multiple decision trees. After the results of all these decision trees are added up, the final result can be obtained. Specifically, the iteration of GBDT can use the forward distribution algorithm (Forward, Stagewise, Algorithm), and the weak learner can use the CART regression tree model. In the GBDT iteration, assuming that the strong learner obtained in the previous iteration is ft-1 (x) and the loss function is L (y, ft-1 (x)), the goal of this round of iteration can be to find a CART The weak learner ht (x) of the regression tree model, so that the loss L (y, ft (x)) = L (y, ft-1 (x)) + ht (x) of this round is the smallest. It can be understood that the decision tree found in this round of iteration is to make the loss of the sample as small as possible. In this embodiment, the Gradient Boosting Decision Number (GBDT) model can be obtained by training as follows: B10: Obtain a sample set for training; the sample set is composed of a preset number of road condition factors and time data pairs. ; For example, the sample set Z shown below: Z = {(x1, y1), (x2, y2), (x3, y3), ..., (xi, yi), ..., (xn, yn) }, Where xi may represent a traffic condition factor corresponding to a travel route from the first geographic position to the i-th second geographic position, and yi may represent time data required to travel the i-th travel route. B20: A gradient boosted decision tree regression algorithm is combined with the sample set to obtain a gradient boosted decision tree model. Generally, the gradient boosting decision tree regression algorithm also needs to configure the maximum number of iterations T and the loss function L. The maximum number of iterations T may be an empirical value set artificially. The loss function L may be a loss function commonly used in GBDT regression calculations in the industry. For example, the mean squared loss function: L (y, f (x)) = (yf (x)) 2, and the absolute loss function: L (y, f (x)) = | yf (x) |, the absolute loss function corresponding to the negative gradient error is sign (yi-f (xi)); the sign function is a sign function. It should be noted that the above-mentioned loss function is only an example, and this embodiment does not limit the specific loss function used.

Specifically, the training process of the gradient boosting decision tree model includes: First, initialize a weak learner as follows:

Next: Perform the iterative calculation of T (maximum number of iterations); each iteration process is as follows:

D10: Calculate a negative gradient based on the data of i = 1, 2, ..., n in the sample set Z. Specifically, the negative gradient can be calculated by the following formula:

Among them, r _ti is the i-th negative gradient, the partial derivative sign, and L (y, f (xi)) is the loss function. That is, each x corresponds to a negative gradient, (xi, r _ti ), i = 1, 2, ..., n.

D20: Fit a CART regression tree based on the calculated negative gradient.

As mentioned earlier, according to the negative gradient (xi, r _ti ), i = 1,2, ..., n; a CART regression tree can be fitted, where the corresponding leaf node area R _tj , j = 1,2 , ..., J, J are the number of leaf nodes of the CART regression tree.

D30: calculate a best fit value according to the leaf area of the CART regression tree.

The best fit value can be calculated by the following formula:

Among them, c _tj is the best fit value, R _tj is the leaf area of the CART regression tree, and j = 1,2, ..., J, J is the number of leaf nodes of the CART regression tree.

D40: Update the strong learner according to the calculated best fit value to obtain the updated strong learner;

Specifically, the updated strong learner is:

D50: Use the updated strong learner as the weak learner for the next iteration, and perform the next iteration; until the maximum number of iterations is reached.

Finally, the expression of the strong learner calculated by the last iteration is:

Determined as the GBDT model. At this point, the GBDT model has been constructed.

E. According to the second geographical location corresponding to the shortest required travel route determined by the decision, it is determined that the investigation task of the case is assigned to the surveyor in the second geographical location.

It can be known from the above-mentioned embodiments that the electronic device proposed in this application first obtains auto insurance case information after receiving a car insurance report request, and the case information includes a first geographic location and a case type of the insurance case; and secondly based on the case type Query the mapping table between the pre-stored case types, survey tasks, and surveyors to determine the survey tasks for the case and the surveyors that match the survey tasks, and obtain the second geographical location of each matching surveyor Position; then determine each travel route between the first geographic location and the second geographic location, and obtain the determined traffic condition factors on each of the travel routes; and enter the obtained traffic condition factors on each of the travel routes into the gradient again The decision tree model is improved to obtain the prediction result data output by the gradient boosted decision tree model, where the prediction result data is the travel route with the shortest time required for decision from the determined travel routes; finally, according to the decision, The second geographic location corresponding to the shortest travel route In the second survey staff geographic distribution of survey tasks of the case. It can eliminate unexpected factors that prevent surveyors from arriving on time according to road conditions factors, improve the accuracy of case handling in a timely manner, and improve customer satisfaction with services.

In addition, the vehicle insurance survey and dispatching program based on the road condition factor of this application according to the different functions implemented by each part can be described by a program module having the same function. Please refer to FIG. 2, which is a schematic diagram of a program module of a vehicle insurance survey and dispatching schedule based on a road condition factor in an embodiment of an electronic device of the present application. In this embodiment, the vehicle insurance survey and dispatching program based on the road condition factor can be divided into an acquisition module 201, a first determination module 202, a second determination module 203, a prediction module 204, and an allocation module according to the functions implemented by its various parts. 205. It can be known from the above description that the program module referred to in the present application refers to a series of computer program instruction segments capable of performing specific functions, and is more suitable than the program to describe the execution process of the automobile insurance survey and dispatch scheduler based on the road condition factor in the electronic device 10. The functions or operation steps implemented by the modules 201-205 are similar to the above, which will not be described in detail here. By way of example, for example:

The obtaining module 201 is configured to obtain auto insurance case information after receiving the auto insurance report request, where the case information includes a first geographic location of the outbreak case and a case type;

The first determining module 202 is configured to query a mapping table between a pre-stored case type, an investigation task, and an inspector based on the case type, to determine the investigation task of the case and each inspector matching the investigation task, and obtain The second geographical location where each surveyor currently matches;

The second determining module 203 is configured to determine each travel route between the first geographical location and the second geographical location, and obtain a road condition factor on each determined travel route;

The prediction module 204 is configured to input the obtained road condition factors on each travel route into a gradient boosting decision tree model to obtain prediction result data output by the gradient boosting decision tree model, where the prediction result data is determined from the travels The route with the shortest decision time required for the route;

The allocation module 205 is configured to determine that the investigation task of the case is assigned to the surveyor in the second geographical location according to the second geographical location corresponding to the shortest required travel route determined by the decision.

In addition, this application also proposes a road condition factor-based vehicle insurance survey and dispatch method based on road condition factors. Please refer to FIG. 3. The road condition factor-based vehicle insurance survey and dispatch method based on road condition factors includes the following: step:

S301. After receiving the auto insurance report request, obtain auto insurance case information, where the case information includes the first geographic location of the insurance case and the case type;

S302: Query the mapping relationship table between the pre-stored case type, the survey task, and the surveyor based on the case type, to determine the survey task of the case and each surveyor matching the survey task, and obtain each matching survey The second geographic location where the employee is currently located;

S303. Determine travel routes between the first geographic location and the second geographic location, respectively, and obtain road condition factors on the determined journey routes.

S304. Input the obtained road condition factors on each travel route into a gradient boosting decision tree model to obtain prediction result data output from the gradient boosting decision tree model, where the prediction result data is from the determined travel routes, The shortest travel route to be decided;

Specifically, the gradient boosting decision tree model is referred to as GBDT (Gradient Boosting Decision Tree) or MART (Multiple Additive Regression Tree), GBT (Gradient Boosting Tree), GBT (Gradient Tree Boosting), or GBRT (Gradient Boosting Regression Tree). In this embodiment, they are collectively referred to as GBDT. The GBDT is an iterative decision tree algorithm. The algorithm can be composed of multiple decision trees. After the results of all these decision trees are added up, the final result can be obtained. Specifically, the iteration of GBDT can use the forward distribution algorithm (Forward, Stagewise, Algorithm), and the weak learner can use the CART regression tree model. In the GBDT iteration, assuming that the strong learner obtained in the previous iteration is ft-1 (x) and the loss function is L (y, ft-1 (x)), the goal of this round of iteration can be to find a CART The weak learner ht (x) of the regression tree model, so that the loss L (y, ft (x)) = L (y, ft-1 (x)) + ht (x) of this round is the smallest. It can be understood that the decision tree found in this round of iteration is to make the loss of the sample as small as possible. In this embodiment, the Gradient Boosting Decision Number (GBDT) model can be trained by: B11: Obtaining a sample set for training; the sample set is composed of a preset number of road condition factors and time data pairs ; For example, the sample set Z shown below: Z = {(x1, y1), (x2, y2), (x3, y3), ..., (xi, yi), ..., (xn, yn) }, Where xi may represent a traffic condition factor corresponding to a travel route from the first geographic position to the i-th second geographic position, and yi may represent time data required to travel the i-th travel route. B21: A gradient boosted decision tree regression algorithm is combined with the sample set to obtain a gradient boosted decision tree model. Generally, the gradient boosting decision tree regression algorithm also needs to configure the maximum number of iterations T and the loss function L. The maximum number of iterations T may be an empirical value set artificially. The loss function L may be a loss function commonly used in GBDT regression calculations in the industry. For example, the mean square loss function: L (y, f (x)) = (yf (x)) 2, and the absolute loss function: L (y, f (x)) = | yf (x) |, the absolute loss function corresponding to the negative gradient error is sign (yi-f (xi)); the sign function is a sign function. It should be noted that the above-mentioned loss function is only an example, and this embodiment does not limit the specific loss function used.

Next: Perform the iterative calculation of T (maximum number of iterations);

The process of each iteration is as follows: D11: According to the data of i = 1, 2, ..., n in the sample set Z, calculate a negative gradient. Specifically, the calculation of the negative gradient can be performed by the following formula:

D21: Fit a CART regression tree based on the calculated negative gradient.

D31: calculate a best fit value according to the leaf area of the CART regression tree.

The best fit value can be calculated by the following formula:

D41: Update the strong learner according to the calculated best fit value to obtain the updated strong learner;

Specifically, the updated strong learner is:

D51: Use the updated strong learner as the weak learner for the next iteration, and perform the next iteration; until the maximum number of iterations is reached.

Finally, the expression of the strong learner calculated from the last iteration

S305. According to the second geographical location corresponding to the shortest required travel route determined by the decision, it is determined that the investigation task of the case is assigned to the surveyor in the second geographical location.

It can be known from the above-mentioned embodiments that the road insurance factor-based automobile insurance survey and dispatch method proposed in the present application first obtains automobile insurance case information after receiving the automobile insurance report request, and the case information includes the first geographic location of the insurance case and the case type; Secondly, based on the case type, query the mapping relationship table between the pre-stored case type and the survey task and the surveyor to determine the survey task of the case and each surveyor matching the survey task, and obtain each matching surveyor. The second geographic location where the current location is currently located; then each travel route between the first geographic location and the second geographic location is determined separately, and the determined traffic condition factors on each of the travel routes are obtained; The traffic condition factor is input to the gradient boosting decision tree model to obtain the prediction result data output by the gradient boosting decision tree model, where the prediction result data is the shortest travel time determined from the determined travel routes. Route; the shortest travel route corresponding to the final decision-making time The second location, determine the allocation survey mission to survey members of the case in the second geographical location. It can exclude accidental factors that prevent investigators from arriving on time according to road conditions factors, improve the timely accuracy of case processing, and improve customer satisfaction with services.

In addition, the present application also proposes a computer-readable storage medium that stores a road insurance factor-based vehicle insurance survey and dispatch program based on road condition factors, and the road condition factor-based vehicle insurance survey and scheduling program is based on When the vehicle risk survey and dispatch program of the road condition factor is executed by the processor, the following operations are performed:

Based on the case type, query the pre-stored mapping table between the case type, the survey task, and the surveyor to determine the survey task of the case and each surveyor matching the survey task, and obtain the matching surveyor's current Second geographical location

According to the second geographical location corresponding to the shortest required travel route determined by the decision, it is determined that the investigation task of the case is assigned to the surveyor in the second geographical location.

The specific implementation of the computer-readable storage medium of the present application is basically the same as the above-mentioned electronic device and various embodiments of the method for vehicle risk survey and dispatch based on road condition factors, and will not be repeated here.

The above-mentioned serial numbers of the embodiments of the present application are merely for description, and do not represent the superiority or inferiority of the embodiments.

Through the description of the above embodiments, those skilled in the art can clearly understand that the methods in the above embodiments can be implemented by means of software plus a necessary universal hardware platform, and of course, also by hardware, but in many cases the former is better. Implementation. Based on such an understanding, the technical solution of this application that is essentially or contributes to the existing technology can be embodied in the form of a software product, which is stored in a storage medium (such as ROM / RAM, magnetic disk, The optical disc) includes several instructions for causing a terminal device (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) to execute the methods described in the embodiments of the present application.

The above are only preferred embodiments of the present application, and thus do not limit the patent scope of the present application. Any equivalent structure or equivalent process transformation made by using the description and drawings of the present application, or directly or indirectly used in other related technical fields Are included in the scope of patent protection of this application.

Claims

An electronic device, wherein the electronic device includes a memory and a processor connected to the memory, the processor is configured to execute a road condition factor-based automobile insurance survey and dispatching program stored on the memory, and the based When the vehicle risk survey and dispatch program of the road condition factor is executed by the processor, the following steps are implemented:

A10. After receiving the auto insurance report request, obtain auto insurance case information, where the case information includes the first geographic location of the case and the type of the case;

A20: Query the mapping table between the pre-stored case types, survey tasks, and surveyors based on the case types to determine the survey tasks of the case and the surveyors that match the survey tasks, and obtain the matching surveys. The second geographic location where the employee is currently located;

A30. Determine travel routes between the first geographic location and the second geographic location, respectively, and obtain road condition factors on the determined travel routes.

A40. Input the obtained road condition factors on each travel route into a gradient boosting decision tree model to obtain prediction result data output from the gradient boosting decision tree model, where the prediction result data is from the determined travel routes, The shortest travel route to be decided;

A50. Determine, according to the second geographic location corresponding to the shortest travel route determined by the decision, to assign the survey task of the case to the surveyor in the second geographic location.
The electronic device according to claim 1, wherein in the step A40, the training process of the gradient boosting decision tree model comprises:

B10: obtaining a sample set for training; the sample set is composed of a preset number of road condition factors and time data pairs;

The sample set: Z: Z = {(x1, y1), (x2, y2), (x3, y3), ..., (xi, yi), ..., (xn, yn)}, where , Xi may represent a traffic condition factor corresponding to a travel route from the first geographic position to the i-th second geographic position, and yi may represent time data required to travel the i-th travel route;

B20: A gradient boosted decision tree regression algorithm is combined with the sample set to obtain a gradient boosted decision tree model.
The electronic device according to claim 2, wherein the step B20 comprises:

C10. Configure the maximum number of iterations t and loss function L of the gradient boost decision tree regression algorithm;

C20. Initialize the weak learner of the loss function L:

The iterative calculation of the maximum number of iterations of the initialized weak learner is performed for t rounds to obtain a gradient boosting decision tree model.
The electronic device according to claim 3, wherein in the step C20, each iterative calculation process for the initialized weak learner comprises:

D10. Substituting the data of i = 1, 2, ..., n in the sample set Z into the formula for calculating the negative gradient, and calculating the negative gradient, the formula for calculating the negative gradient is:

Where r ti is the ith negative gradient,
Is the partial derivative sign, and L (y, f (xi)) is the loss function;

D20: fit a CART regression tree according to the calculated negative gradient;

D30: Substituting the leaf area of the CART regression tree into a formula for calculating a best fit value, the formula for calculating the best fit value is:

Among them, c tj is the best fit value, R tj is the leaf area of the CART regression tree, and j = 1, 2, ..., J, j is the number of leaf nodes of the CART regression tree.

D40: Update the weak learner to obtain a strong learner according to the calculated best fit value. The strong learner is:

D50: Use the updated strong learner as the weak learner in the next iteration, and repeat the above steps D10, D20, and D30 to perform the next iteration until the maximum number of iterations t is reached.
The electronic device according to claim 1, wherein the road condition factor includes road subgrade, road surface, surrounding buildings, and road traffic conditions of the travel route.
The electronic device according to claim 2, wherein the road condition factor includes road subgrade, pavement, road surrounding buildings, and road traffic conditions of the travel route.
The electronic device according to claim 3, wherein the road condition factor includes road subgrade, road surface, surrounding buildings, and road traffic conditions of the travel route.
The electronic device according to claim 4, wherein the road condition factor includes road subgrade, road surface, surrounding buildings, and road traffic conditions of the travel route.
A method for vehicle insurance survey and dispatch based on road condition factors, characterized in that the method includes the following steps:

S10. After receiving the auto insurance report request, obtain auto insurance case information, where the case information includes the first geographic location of the outbreak case and the case type;

S20: Query the mapping table between the pre-stored case type, the survey task, and the surveyor based on the case type to determine the survey task of the case and each surveyor matching the survey task, and obtain each matching survey The second geographic location where the employee is currently located;

S30. Determine travel routes between the first geographic location and the second geographic location, respectively, and obtain road condition factors on the determined travel routes.

S40. Input the obtained road condition factors on each travel route into a gradient promotion decision tree model to obtain prediction result data output by the gradient promotion decision tree model, where the prediction result data is from the determined travel routes, The shortest travel route to be decided;

S50. Determine, according to the second geographical location corresponding to the shortest required travel route determined by the decision, to allocate the investigation task of the case to the surveyor in the second geographical location.
The method of claim 9, wherein in step S40, the training process of the gradient boosting decision tree model comprises:

B11: obtaining a sample set for training; the sample set is composed of a preset number of road condition factors and time data pairs;

The sample set: Z: Z = {(x1, y1), (x2, y2), (x3, y3), ..., (xi, yi), ..., (xn, yn)}, where , Xi may represent a traffic condition factor corresponding to a travel route from the first geographic position to the i-th second geographic position, and yi may represent time data required to travel the i-th travel route;

B21: A gradient boosted decision tree regression algorithm is combined with the sample set to obtain a gradient boosted decision tree model.
The method according to claim 10, wherein the step B21 comprises:

C11. Configure the maximum iterations t and loss function L of the gradient-enhancing decision tree regression algorithm;

C21. Initialize the weak learner of the loss function L:

The iterative calculation of the maximum number of iterations of the initialized weak learner is performed for t rounds to obtain a gradient boosting decision tree model.
The method of claim 11, wherein in the step C21, each iterative calculation process of the initialized weak learner comprises:

D11. Substituting the data of i = 1, 2, ..., n in the sample set Z into the formula for calculating the negative gradient, and calculating the negative gradient, the formula for calculating the negative gradient is:

Where r ti is the ith negative gradient,
Is the partial derivative sign, and L (y, f (xi)) is the loss function;

D21: Fit a CART regression tree based on the calculated negative gradient;

D31: Substituting the leaf area of the CART regression tree into a formula for calculating a best fit value, the formula for calculating the best fit value is:

Among them, c tj is the best fit value, R tj is the leaf area of the CART regression tree, and j = 1,2, ..., J, J is the number of leaf nodes of the CART regression tree.

D41: Update the weak learner to obtain a strong learner according to the calculated best fit value. The strong learner is:

D51: Use the updated strong learner as the weak learner for the next iteration, and repeat the above steps D11, D21, and D31 to perform the next iteration until the maximum number of iterations t is reached
The method according to claim 9, wherein the road condition factor includes road subgrade, road surface, surrounding buildings, and road traffic conditions of the travel route.
The method of claim 10, wherein the road condition factor comprises road subgrade, pavement, road surrounding buildings, and road traffic conditions of the travel route.
The method according to claim 11, wherein the road condition factor comprises road subgrade, road surface, surrounding buildings, and road traffic conditions of the travel route.
The method according to claim 12, wherein the road condition factor comprises road subgrade, pavement, surrounding buildings, and road traffic conditions of the travel route.
A computer-readable storage medium stores a road insurance factor-based vehicle risk survey and dispatch program based on the road condition factor, and the road condition factor-based vehicle risk survey and dispatch program can be executed by at least one processor to enable the at least one The processor performs the following steps:

After receiving the auto insurance report request, obtain auto insurance case information, where the case information includes the first geographic location of the outbreak case and the case type;

Based on the case type, query the pre-stored mapping table between the case type, the survey task, and the surveyor to determine the survey task of the case and each surveyor matching the survey task, and obtain the matching surveyor's current Second geographical location

Determining travel routes between the first geographic location and the second geographic location, respectively, and acquiring traffic condition factors on the determined travel routes;

The obtained traffic condition factors on each travel route are input to a gradient boosting decision tree model to obtain prediction result data output by the gradient boosting decision tree model, where the prediction result data is determined from the determined travel routes The shortest travel route required;

According to the second geographic location corresponding to the shortest required travel route determined by the decision, it is determined that the investigation task of the case is assigned to the surveyor in the second geographic location.
The computer-readable storage medium of claim 17, wherein the training process of the gradient boosting decision tree model comprises:

Obtaining a sample set for training; the sample set is composed of a preset number of road condition factors and time data pairs;

The sample set: Z: Z = {(x1, y1), (x2, y2), (x3, y3), ..., (xi, yi), ..., (xn, yn)}, where , Xi may represent a traffic condition factor corresponding to a travel route from the first geographic position to the i-th second geographic position, and yi may represent time data required to travel the i-th travel route;

A gradient boosted decision tree regression algorithm is combined with the sample set to train a gradient boosted decision tree model.
The computer-readable storage medium of claim 18, wherein the step of obtaining a gradient-boosted decision tree model based on the gradient-boosted decision tree regression algorithm combined with training of the sample set comprises:

Configure the maximum iterations t and loss function L of the gradient boosting decision tree regression algorithm;

Initialize the weak learner of the loss function L:

The iterative calculation of the maximum number of iterations of the initialized weak learner is performed for t rounds to obtain a gradient boosting decision tree model.
The computer-readable storage medium of claim 19, wherein the process of performing an iterative calculation on the initialized weak learner each time comprises:

D10. Substituting the data of i = 1, 2, ..., n in the sample set Z into the formula for calculating the negative gradient, and calculating the negative gradient, the formula for calculating the negative gradient is:

Where r ti is the ith negative gradient,
Is the partial derivative sign, and L (y, f (xi)) is the loss function;

D20: fit a CART regression tree according to the calculated negative gradient;

D30: Substituting the leaf area of the CART regression tree into a formula for calculating a best fit value, the formula for calculating the best fit value is:

Among them, c tj is the best fit value, R tj is the leaf area of the CART regression tree, and j = 1, 2, ..., J, j is the number of leaf nodes of the CART regression tree.

D40: Update the weak learner to obtain a strong learner according to the calculated best fit value. The strong learner is:

D50: Use the updated strong learner as the weak learner in the next iteration, and repeat the above steps D10, D20, and D30 to perform the next iteration until the maximum number of iterations t is reached.