CN115018658A

CN115018658A - Loss price early warning method, loss price early warning device, loss price early warning equipment, storage medium and program product

Info

Publication number: CN115018658A
Application number: CN202210947522.7A
Authority: CN
Inventors: 黄玉华; 马丹雄; 黄锐; 陈中乾
Original assignee: Taiping General Insurance Co ltd; Taiping Financial Technology Services Shanghai Co Ltd Shenzhen Branch
Current assignee: Taiping General Insurance Co ltd; Taiping Financial Technology Services Shanghai Co Ltd Shenzhen Branch
Priority date: 2022-08-09
Filing date: 2022-08-09
Publication date: 2022-09-06

Abstract

The application relates to a loss assessment price early warning method, a loss assessment price early warning device, loss assessment price early warning equipment, loss assessment price early warning media and products. The method comprises the following steps: obtaining damage assessment information, wherein the damage assessment information comprises a current damage assessment price and current damage assessment data; inputting the current damage assessment data and the current damage assessment price into an early warning model obtained by pre-training for processing so as to obtain first early warning information of the current damage assessment price; acquiring standard prices of accessories calculated in advance according to statistics; obtaining second early warning information according to the standard price of the accessories and the current damage price; and obtaining target early warning information of the loss assessment price according to the first early warning information and the second early warning information. The method is adopted to judge the current loss assessment price according to different modes, so that the accuracy is improved, manual communication is not needed, and the labor input is reduced.

Description

Loss price early warning method, loss price early warning device, loss price early warning equipment, storage medium and program product

Technical Field

The application relates to the technical field of artificial intelligence, in particular to a method, a device, equipment, a medium and a product for early warning loss price.

Background

In the stages of loss assessment, loss checking and the like of vehicle insurance claims, the price loss of the vehicle loss needs to be estimated; for the parts needing to be replaced, the actual price of the parts needs to be evaluated as much as possible. The price of accessories of vehicles of different brands and models is different due to the influence of the quality and the area of the accessories, and the price of the accessories is not uniform.

Conventionally, a loss assessment service person inquires a local accessory dealer about a price, and maintains the price as a reference price in a claim settlement system by region, sets a discount according to the degree of cooperation with a repair shop, and calculates the price and discount in the claim settlement system. And negotiating the final loss assessment price with a repair shop, and actually settling the claim at the negotiated price.

However, in the current manual inquiry, the price quoted by different accessory providers is different, and the labor cost is higher.

Disclosure of Invention

In view of the above, it is necessary to provide a damage-assessment price early-warning method, apparatus, device, medium, and product that can reduce labor costs.

In a first aspect, the present application provides a loss assessment price early warning method, including:

obtaining damage assessment information, wherein the damage assessment information comprises a current damage assessment price and current damage assessment data;

inputting the current damage assessment data and the current damage assessment price into an early warning model obtained by pre-training for processing so as to obtain first early warning information of the current damage assessment price;

acquiring standard prices of accessories calculated in advance according to statistics;

obtaining second early warning information according to the standard price of the accessories and the current damage price;

and obtaining target early warning information of the loss assessment price according to the first early warning information and the second early warning information.

In one embodiment, the method further comprises:

acquiring historical claim settlement data, wherein the historical claim settlement data comprises historical damage assessment data and historical damage assessment prices;

performing feature vectorization on the historical damage assessment data to obtain a sample feature vector;

and training according to the sample feature vector and the historical damage assessment price to obtain an early warning model.

In one embodiment, before performing feature vectorization on the historical damage assessment data to obtain a sample feature vector, the method further includes:

dividing the historical damage assessment data into at least one class, and selecting a current sample center in each class;

calculating the current distance between the historical damage assessment data in each class and the current sample center;

calculating to obtain a new sample center of each type according to the first statistical value of the current distance;

judging the difference value between the current sample center and the new sample center;

when the difference value is smaller than a preset threshold value, taking the new sample center as a target sample center, otherwise, taking the new sample center as a current sample center, and continuously calculating the current distance between the historical damage assessment data in each class and the current sample center until the difference value is smaller than the preset threshold value;

and carrying out standardization processing on the historical damage assessment data according to the target sample center.

In one embodiment, the normalizing the historical damage assessment data according to the target sample center includes:

calculating a first distance from the historical damage assessment data to the center of the target sample; calculating a second statistical value of the first distance to obtain a reference distance;

and carrying out standardization processing on the historical damage assessment data according to the first distance and the reference distance.

In one embodiment, the inputting the current damage assessment data and the current damage assessment price into a pre-trained early warning model for processing to obtain first early warning information of the current damage assessment price includes:

inputting the current damage assessment data into an early warning model obtained by pre-training to obtain a predicted price;

when the predicted price is larger than or equal to the current loss price, generating first early warning information that the current loss price is normal;

and when the predicted price is smaller than the current loss price, generating first early warning information that the current loss price is abnormal.

In one embodiment, the method further comprises:

and calculating a third statistical value of the historical damage assessment data in each class, and taking the third statistical value as the standard price of the accessories.

In one embodiment, the obtaining second warning information according to the standard price of the accessories and the current damage-assessment price includes:

calculating discrete coefficients of historical damage assessment data in each class, and calculating to obtain a reference threshold according to the discrete coefficients;

and obtaining second early warning information according to the reference threshold and the current damage assessment price.

In one embodiment, the calculating a reference threshold according to the discrete coefficient includes:

when the discrete coefficient is smaller than or equal to a threshold ratio, calculating to obtain upper and lower limits of a reference threshold according to the median and the standard deviation of the historical damage assessment data;

and when the discrete coefficient is larger than the threshold ratio, calculating to obtain the upper limit and the lower limit of the reference threshold according to the average value and the median of the historical damage assessment data and the threshold ratio.

In one embodiment, the obtaining target early warning information of the loss assessment price according to the first early warning information and the second early warning information includes:

when at least one of the first early warning information and the second early warning information is that the current damage assessment price is abnormal, outputting target early warning information of the abnormal damage assessment price;

and when the first early warning information and the second early warning information are not the current loss assessment price abnormity, outputting the target early warning information with normal loss assessment price.

In one embodiment, after the outputting of the target warning information of the abnormal current loss assessment price, the method includes:

receiving an adjustment instruction for the current damage-fixing price;

and adjusting the current damage assessment price according to the adjusting instruction, wherein the adjusted current damage assessment price is used for updating at least one of historical claims data, the early warning model and the standard price of the accessories.

In a second aspect, the present application provides a loss assessment price early warning device, the device includes:

the data acquisition module is used for acquiring loss assessment information, wherein the loss assessment information comprises a current loss assessment price and current loss assessment data;

the first early warning module is used for inputting the current damage assessment data and the current damage assessment price into an early warning model obtained by pre-training for processing so as to obtain first early warning information of the current damage assessment price;

the accessory standard price acquisition module is used for acquiring an accessory standard price which is calculated in advance according to statistics;

the second early warning module is used for obtaining second early warning information according to the standard price of the accessories and the current damage price;

and the target early warning module is used for obtaining the target early warning information of the damage assessment price according to the first early warning information and the second early warning information.

In a third aspect, the present application provides a computer device comprising a memory storing a computer program and a processor implementing the steps of the method in any of the above embodiments when the processor executes the computer program.

In a fourth aspect, the present application provides a computer-readable storage medium having stored thereon a computer program which, when executed by a processor, performs the steps of the method in any of the above-described embodiments.

In a fifth aspect, the present application provides a computer program product comprising a computer program which, when executed by a processor, performs the steps of the method in any of the embodiments described above.

According to the loss assessment price early warning method, device, equipment, medium and product, after loss assessment information is obtained, first early warning information is obtained by processing according to current loss assessment data and loss assessment price through an early warning model, second early warning information is obtained according to standard price of accessories and the current loss assessment price obtained through statistical calculation, and finally target early warning information is obtained by processing according to the first early warning information and the second early warning information.

Drawings

FIG. 1 is a diagram of an application environment of a damage-assessment price early warning method in one embodiment;

FIG. 2 is a schematic flow chart illustrating a loss assessment price warning method according to an embodiment;

FIG. 3 is a schematic flow chart of the pretreatment step in one embodiment;

FIG. 4 is a schematic flow chart illustrating a loss assessment price warning method according to another embodiment;

FIG. 5 is a block diagram of a loss assessment price warning device in one embodiment;

FIG. 6 is a diagram illustrating an internal structure of a computer device according to an embodiment.

Detailed Description

In order to make the objects, technical solutions and advantages of the present application more apparent, the present application is described in further detail below with reference to the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative of the present application and are not intended to limit the present application.

The loss assessment price early warning method provided by the embodiment of the application can be applied to the application environment shown in the figure 1. Wherein the terminal 102 communicates with the server 104 via a network. The data storage system may store data that the server 104 needs to process. The data storage system may be integrated on the server 104, or may be located on the cloud or other network server.

The terminal 102 may upload damage assessment information to the server 104, where the damage assessment information includes a current damage assessment price and current damage assessment data; the server 104 inputs the current damage assessment data and the current damage assessment price into a pre-trained early warning model for processing to obtain first early warning information of the current damage assessment price; on the other hand, the server 104 acquires a standard price of the accessory calculated in advance according to statistics; obtaining second early warning information according to the standard price of the accessories and the current damage-assessment price; and finally, the server 104 obtains target early warning information of the loss price according to the first early warning information and the second early warning information. After the damage assessment information is obtained, first early warning information is obtained by processing according to current damage assessment data and a damage assessment price through an early warning model, then second early warning information is obtained according to a standard price of accessories and the current damage assessment price which are obtained through statistical calculation, and finally target early warning information is obtained by processing according to the first early warning information and the second early warning information.

The terminal 102 may be, but not limited to, various personal computers, notebook computers, smart phones, tablet computers, internet of things devices and portable wearable devices, and the internet of things devices may be smart speakers, smart televisions, smart air conditioners, smart car-mounted devices, and the like. The portable wearable device can be a smart watch, a smart bracelet, a head-mounted device, and the like. The server 104 may be implemented as a stand-alone server or as a server cluster comprised of multiple servers.

In one embodiment, as shown in fig. 2, a loss assessment price early warning method is provided, which is described by taking the method as an example applied to the server in fig. 1, and includes the following steps:

s202: and obtaining the damage assessment information, wherein the damage assessment information comprises the current damage assessment price and the current damage assessment data.

Specifically, the damage assessment information refers to case information of each damage assessment core damage, and includes a predetermined current damage assessment price and other information of the case, the other information of the case includes current damage assessment data and the like, and the current damage assessment data includes characteristics of license plates, vehicle types, accessories, regions, historical reference prices, actual accessory prices, repair plants, repair plant cooperation models and the like.

S204: and inputting the current damage assessment data and the current damage assessment price into an early warning model obtained by pre-training for processing so as to obtain first early warning information of the current damage assessment price.

Specifically, the early warning model is obtained by pre-training, wherein the early warning model is used for calculating and obtaining first early warning information according to current damage assessment data. Specifically, the early warning model may be obtained through XGboost model training according to historical claims data, and in other embodiments, other model algorithms may be used for processing. The historical claim settlement data comprises the characteristics of license plates, vehicle types, accessories, regions, historical reference prices, actual accessory prices, repair plants, repair plant cooperation models and the like, so that the actual accessory prices are used as reference outputs, other characteristics are used as inputs, real-time outputs are obtained through model calculation, and loss functions are calculated according to the reference outputs and the real-time outputs so as to train the early warning models.

Preferably, the server inputs the current damage assessment data into a pre-trained early warning model for processing to obtain a target price corresponding to the current damage assessment data, and first early warning information can be generated according to the target price and the current damage assessment price.

S206: and acquiring standard prices of the accessories, which are calculated in advance according to statistics.

Specifically, the standard price of the accessory is calculated in advance according to a statistical mode, namely, historical claim settlement data is processed statistically to obtain the standard price of the accessory.

Specifically, the server may perform statistical processing according to the historical claim data, for example, calculate a median of all the historical claim data as the standard price of the accessory, or calculate a range according to the median as the standard price of the accessory.

In other embodiments, the server may also perform a clustering process in advance according to the historical claims data, for example, according to the quality information of the accessories, so as to obtain a plurality of classifications, for example, 4S original shop parts, OEM parts, brand parts, and general parts. The server calculates the historical claim settlement data in each category to obtain statistical information, such as a median, and the median is used as the standard price of the accessories of the category.

S208: and obtaining second early warning information according to the standard price of the accessories and the current damage-assessment price.

Specifically, the server may compare the standard price of the accessory with the current damage price to obtain the second warning information, which may be specifically referred to below.

S210: and obtaining target early warning information of the damage assessment price according to the first early warning information and the second early warning information.

In one embodiment, when at least one of the first early warning information and the second early warning information is that the current loss assessment price is abnormal, target early warning information with abnormal loss assessment price is output; and when the first early warning information and the second early warning information are not abnormal to the current damage assessment price, outputting target early warning information with normal damage assessment price.

In practical application, the early warning model is taken as an XGboost classification model for example, the first early warning information of the XGboost classification model is L, and the second early warning information of statistical calculation is M; if the conclusion of L and M is normal, the price of the verification loss is reasonable, if one of the conclusion is abnormal, the target early warning result is abnormal, and the business is required to verify, namely the business communicates to re-determine the price.

According to the loss assessment price early warning method, after loss assessment information is obtained, first early warning information is obtained by processing according to current loss assessment data and loss assessment price through an early warning model, second early warning information is obtained according to standard price of accessories and the current loss assessment price which are obtained through statistical calculation, and finally target early warning information is obtained by processing according to the first early warning information and the second early warning information.

In one embodiment, the loss assessment price early warning method further includes a training process of an early warning model, where the training process of the early warning model includes: acquiring historical claim settlement data, wherein the historical claim settlement data comprises historical damage assessment data and historical damage assessment prices; performing feature vectorization on the historical damage assessment data to obtain a sample feature vector; and training according to the sample feature vector and the historical damage assessment price to obtain an early warning model.

Specifically, in the embodiment, the reference price of the accessory is calculated by using the XGBoost classification algorithm based on the historical claim settlement data and using the vehicle type, the vehicle series, the accessory number, the area, the accessory channel, and the accessory quality as classification points.

The main flow comprises collecting historical claims data of the automobile insurance; cleaning the historical claims data of the automobile insurance, and screening abnormal values by using a clustering algorithm; obtaining the characteristics of license plates, vehicle types, accessories, regions, historical reference prices, actual accessory prices, repair plants, repair plant cooperation models and the like from historical claims; taking the characteristic value extracted from the historical claims data of the automobile insurance as a training data set to train an early warning model; calculating and early warning the price of the damaged vehicle parts by using an early warning model; and adjusting the damage price of the vehicle accessories according to the early warning result.

Specifically, the historical claim settlement data can be obtained from a vehicle insurance claim settlement system, and the historical claim settlement data includes policy information, case information, accessory information and repair shop information, wherein the accessory information must include data of vehicle brand, vehicle type, vehicle accessory and serial number, location, price of actual claim settlement, time and the like.

The server vectorizes the characteristics of the historical damage assessment data to obtain a sample characteristic vector, wherein the sample characteristic vector can be obtained by One-hot coding, for example, the processed data is vectorized by the characteristics, and fields such as a vehicle type, a fitting number, a region and the like are vectorized by the One-hot coding.

And finally, the server inputs the vectorized sample feature vector into an early warning model to obtain a reference price, then compares the reference price with an actual loss assessment price to obtain a loss function, calculates according to the loss function to train the early warning model, and when the loss function meets requirements, for example, the loss function is smaller than a certain value or the size of the loss function does not change any more, the training is finished to obtain the early warning model.

In the embodiment, the processed data is subjected to feature vectorization, fields such as vehicle types, accessories, accessory numbers, regions and the like are subjected to vectorization through One-hot codes, feature values extracted from vehicle insurance historical claims data are used as training data set training early warning models, and XGB OST classification algorithms are adopted for training the models. The damage price of the training sample after abnormal data are removed is regarded as a reasonable price, data of different areas, different channels, different vehicle types and different accessories are classified into one type through classification, and the price of the data is the actual price of the accessory.

In one embodiment, referring to fig. 3, before performing feature vectorization on historical damage assessment data to obtain a sample feature vector, the method further includes a step of preprocessing the early warning model, where the preprocessing step includes:

s302: and dividing the historical damage assessment data into at least one class, and selecting the current sample center in each class.

Specifically, the historical damage data may be divided into multiple categories according to different quality of the parts, for example, the historical damage data is divided into 4S original shop parts, OEM parts, brand parts, and general parts according to the quality of the parts, and then one of the historical damage data in each category is randomly selected as the current sample center.

In the actual processing, the accessories have different qualities and actual prices, and the accessories are divided into 4 classes according to business experiences. And randomly selecting 4 points as the current sample center.

S304: and calculating the current distance between the historical damage assessment data in each class and the current sample center.

S306: and calculating a new sample center of each type according to the first statistical value of the current distance.

Specifically, the calculation of the current distance may be a vector distance, and the calculation of the first statistical value may be a calculation of a mean value.

That is, the server first calculates the distance of the remaining samples from the cluster center and labels each sample as the closest category to the 4 cluster centers; the mean of the sample points in each cluster was recalculated and taken as the new 4 cluster centers.

S308: and judging the difference value between the current sample center and the new sample center.

S310: and when the difference value is smaller than the preset threshold value, taking the new sample center as a target sample center, otherwise, taking the new sample center as a current sample center, and continuously calculating the current distance between the historical damage assessment data in each class and the current sample center until the difference value is smaller than the preset threshold value.

Specifically, the server repeats the above steps until the difference between the current sample center and the new sample center satisfies a condition, for example, the change of the sample center tends to be stable, thereby finally forming a plurality of classifications to complete the clustering.

S312: and carrying out standardization processing on the historical damage assessment data according to the target sample center.

In one embodiment, the normalizing the historical damage assessment data according to the target sample center includes: calculating a first distance from the historical damage assessment data to the center of the target sample; calculating a second statistical value of the first distance to obtain a reference distance; and carrying out standardization processing on the historical damage assessment data according to the first distance and the reference distance.

Specifically, the server calculates the average value of the distance from each point in the cluster to the center of the cluster based on the clustering result; calculating the ratio R of the distance D from each sample to the central point and the average distance of the central point; if (R-D)/R >0.2, the data is considered abnormal data, and the data is removed from the training sample set. In other embodiments, a value other than 0.2 may be selected, and is not particularly limited herein.

It should be noted that, in the historical claims data, the actual execution prices of the same accessories in the same region are not consistent, the prices of the accessories with different qualities are different, and the actual execution prices of the accessories with the same quality are different due to different quotations of repair shops or other reasons, but most of the actual execution prices of the accessories are within a range, and the lower data and the higher data are abnormal data, and in order to reduce the influence of the abnormal data on the prediction result, the data need to be removed. And identifying abnormal data by using a K-means clustering algorithm. And the K-means clustering algorithm calculates and selects a plurality of central points, and continuously calculates the distance between each sample point and the cluster center until convergence. Specific clustering methods can be found above.

In one embodiment, inputting the current damage assessment data and the current damage assessment price into an early warning model obtained by pre-training to be processed so as to obtain first early warning information of the current damage assessment price, where the method includes: inputting the current damage assessment data into an early warning model obtained by pre-training to obtain a predicted price; when the predicted price is larger than or equal to the current loss assessment price, generating first early warning information that the current loss assessment price is normal; and when the predicted price is smaller than the current loss assessment price, generating first early warning information of abnormal current loss assessment price.

Specifically, the server predicts whether the core loss price is abnormal or not through a model: and performing One-hot coding on the current claim settlement data to be predicted to perform vectorization, and inputting the data into a prediction model to obtain the predicted price of the accessory. The predicted price is compared to the actual check loss price. The predicted price is P and the actual core loss price is R. If P > = R, the price of the fixed core loss is normal, and if P < R, the price of the fixed core loss is higher and is abnormal.

In one embodiment, the method for early warning of damage assessment price may further include calculating a third statistical value of historical damage assessment data in each class, and using the third statistical value as a standard price of the accessory.

Specifically, the server divides the data into 4 classes according to the quality of the accessories by using a KMeans clustering algorithm (the calculation method is the same as that of abnormal value processing), wherein the 4S classes are original manufacturer parts, OEM parts, brand parts and general parts of the shop. And calculating corresponding median of the data in different classifications, wherein the data is the standard price in the quality data, namely the standard price of the accessory.

In one embodiment, the obtaining of the second warning information according to the standard price of the accessories and the current damage-assessment price comprises: calculating discrete coefficients of the historical damage assessment data in each class, and calculating to obtain a reference threshold according to the discrete coefficients; and obtaining second early warning information according to the reference threshold and the current loss assessment price.

In one embodiment, the calculating the reference threshold according to the discrete coefficient includes: when the discrete coefficient is smaller than or equal to the threshold ratio, calculating to obtain the upper limit and the lower limit of a reference threshold according to the median and the standard deviation of the historical damage assessment data; and when the discrete coefficient is larger than the threshold ratio, calculating to obtain the upper limit and the lower limit of the reference threshold according to the average value, the median and the threshold ratio of the historical damage assessment data.

Specifically, the price of the vehicle accessory after damage assessment is calculated and early-warned by using an early-warning model. And calculating the case information of each damage assessment core damage through an early warning model, and predicting whether the price of the current damage assessment accessory exceeds a normal value. Calculating the upper and lower threshold values of the current accessory price through discrete coefficients, wherein the calculation formula is as follows:

coefficient of dispersion = standard deviation/mean

The server sets a threshold ratio R, when | discrete coefficient | < R, the lower limit of the early warning value = the median-standard deviation, and the upper limit of the early warning value = the median + standard deviation. When the | dispersion coefficient | is larger than R and the average value is larger than or equal to 0, the lower warning value limit = median × (1-R) and the upper warning value limit = median × (1 + R). When | dispersion coefficient | > R, and when the average value <0, the lower warning value limit = median × (1 + R), and the upper warning value limit = median × (1-R).

And if the fittings of the model early warning also exceed the upper limit of the early warning value calculated by the discrete coefficient, the damage-assessment price of the current fittings is higher, and if the fittings do not exceed the upper limit of the early warning value, the fittings are priced normally.

In one embodiment, after outputting the target early warning information of the abnormal current loss assessment price, the method includes: receiving an adjustment instruction aiming at the current loss assessment price; and adjusting the current damage assessment price according to the adjusting instruction, wherein the adjusted current damage assessment price is used for updating at least one of historical claim settlement data, the early warning model and the standard price of the accessory.

Specifically, the price in the historical claim settlement data is adjusted according to the verification result: and inputting the result verified by the service personnel into the historical claim settlement data, updating the actual reasonable price of the accessory into the historical claim settlement data, and retraining the model by using the data to correct the error of the model.

Specifically, referring to fig. 4, the server acquires historical claim settlement data, and then performs abnormal value processing on the historical claim settlement data, so that the server performs processing through two threads, wherein the first thread trains a model through an XGBoost classification algorithm, so that whether the price of the core damage is abnormal is predicted through the model, and the other thread calculates the median of the accessory, so that whether the price of the core damage is abnormal is predicted. And the server fuses the prediction result, if the prediction result is abnormal, the service verifies the prediction risk, and adjusts the price in the historical claim settlement data according to the verification result so as to adjust the model and the median.

It should be understood that, although the steps in the flowcharts related to the embodiments are shown in sequence as indicated by the arrows, the steps are not necessarily executed in sequence as indicated by the arrows. The steps are not performed in the exact order shown and described, and may be performed in other orders, unless explicitly stated otherwise. Moreover, at least a part of the steps in the flowcharts related to the above embodiments may include multiple steps or multiple stages, which are not necessarily performed at the same time, but may be performed at different times, and the order of performing the steps or stages is not necessarily sequential, but may be performed alternately or alternately with other steps or at least a part of the steps or stages in other steps.

Based on the same inventive concept, the embodiment of the application also provides a loss assessment price early warning device for realizing the loss assessment price early warning method. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the method, so the specific limitations in one or more loss assessment price early warning device embodiments provided below can be referred to the limitations on the loss assessment price early warning method in the above, and details are not repeated here.

In one embodiment, as shown in fig. 5, there is provided a loss price early warning apparatus including: data acquisition module 501, first early warning module 502, accessory standard price acquisition module 503, second early warning module 504 and target early warning module 505, wherein:

the data acquisition module 501 is configured to acquire loss assessment information, where the loss assessment information includes a current loss assessment price and current loss assessment data;

the first early warning module 502 is configured to input the current damage assessment data and the current damage assessment price into an early warning model obtained through pre-training for processing, so as to obtain first early warning information of the current damage assessment price;

the accessory standard price obtaining module 503 is configured to obtain an accessory standard price calculated in advance according to statistics;

the second early warning module 504 is configured to obtain second early warning information according to the standard price of the accessory and the current damage assessment price;

and the target early warning module 505 is configured to obtain target early warning information of the loss assessment price according to the first early warning information and the second early warning information.

In one embodiment, the loss assessment price early warning device further includes:

the claim data acquisition module is used for acquiring historical claim data, wherein the historical claim data comprises historical damage assessment data and historical damage assessment prices;

the vectorization module is used for carrying out feature vectorization on the historical damage assessment data to obtain a sample feature vector;

and the training module is used for training according to the sample characteristic vector and the historical damage assessment price to obtain an early warning model.

the current sample center calculation module is used for dividing the historical damage assessment data into at least one class and selecting the current sample center in each class;

the current distance calculation module is used for calculating the current distance between the historical damage assessment data in each class and the current sample center;

the new sample center calculating module is used for calculating to obtain a new sample center of each type according to the first statistical value of the current distance;

the judging module is used for judging the difference value between the current sample center and the new sample center;

the updating module is used for taking the new sample center as a target sample center when the difference value is smaller than a preset threshold value, otherwise, taking the new sample center as a current sample center, and continuously calculating the current distance between the historical damage assessment data in each class and the current sample center until the difference value is smaller than the preset threshold value;

and the standardization processing module is used for carrying out standardization processing on the historical damage assessment data according to the target sample center.

In one embodiment, the normalization module includes:

the distance calculation unit is used for calculating a first distance from the historical damage assessment data to the center of the target sample; calculating a second statistical value of the first distance to obtain a reference distance;

and the standardization processing unit is used for carrying out standardization processing on the historical damage assessment data according to the first distance and the reference distance.

In one embodiment, the first warning module 502 includes:

the model processing unit is used for inputting the current damage assessment data into an early warning model obtained by pre-training to obtain a predicted price;

the first early warning unit is used for generating first early warning information that the current damage assessment price is normal when the predicted price is larger than or equal to the current damage assessment price; and when the predicted price is smaller than the current damage assessment price, generating first early warning information of abnormal current damage assessment price.

and the standard price calculating module is used for calculating a third statistical value of the historical damage assessment data in each class and taking the third statistical value as the standard price of the accessory.

In one embodiment, the second warning module 504 includes:

the reference threshold value calculating unit is used for calculating discrete coefficients of the historical damage assessment data in each class and calculating a reference threshold value according to the discrete coefficients;

and the second early warning unit is used for obtaining second early warning information according to the reference threshold and the current damage assessment price.

In one embodiment, the reference threshold calculation unit is configured to calculate, when the discrete coefficient is less than or equal to the threshold ratio, an upper limit and a lower limit of the reference threshold according to a median and a standard deviation of the historical damage assessment data; and when the discrete coefficient is larger than the threshold ratio, calculating to obtain the upper limit and the lower limit of the reference threshold according to the average value, the median and the threshold ratio of the historical damage assessment data.

In one embodiment, the reference threshold calculation unit is configured to output target early warning information with abnormal damage assessment price when at least one of the first early warning information and the second early warning information is that the current damage assessment price is abnormal; and when the first early warning information and the second early warning information are not abnormal in the current loss assessment price, outputting target early warning information with normal loss assessment price.

In one embodiment, the damage-assessment price early-warning device includes:

the receiving module is used for receiving an adjusting instruction aiming at the current loss assessment price;

and the adjusting module is used for adjusting the current damage assessment price according to the adjusting instruction, and the adjusted current damage assessment price is used for updating at least one of historical claim settlement data, the early warning model and the standard price of the accessory.

All or part of the modules in the loss price early warning device can be realized by software, hardware and the combination thereof. The modules can be embedded in a hardware form or independent from a processor in the computer device, and can also be stored in a memory in the computer device in a software form, so that the processor can call and execute operations corresponding to the modules.

In one embodiment, a computer device is provided, which may be a server, and its internal structure diagram may be as shown in fig. 6. The computer device includes a processor, a memory, and a network interface connected by a system bus. Wherein the processor of the computer device is configured to provide computing and control capabilities. The memory of the computer device includes a non-volatile storage medium and an internal memory. The non-volatile storage medium stores an operating system, a computer program, and a database. The internal memory provides an environment for the operation of an operating system and computer programs in the non-volatile storage medium. The database of the computer device is used for storing current damage assessment information, early warning models and the like. The network interface of the computer device is used for communicating with an external terminal through a network connection. The computer program is executed by a processor to implement a loss price warning method.

Those skilled in the art will appreciate that the architecture shown in fig. 6 is merely a block diagram of some of the structures associated with the disclosed aspects and is not intended to limit the computing devices to which the disclosed aspects apply, as particular computing devices may include more or less components than those shown, or may combine certain components, or have a different arrangement of components.

In one embodiment, a computer device is provided, comprising a memory and a processor, the memory having a computer program stored therein, the processor implementing the following steps when executing the computer program: obtaining damage assessment information, wherein the damage assessment information comprises a current damage assessment price and current damage assessment data; inputting the current damage assessment data and the current damage assessment price into an early warning model obtained by pre-training for processing so as to obtain first early warning information of the current damage assessment price; acquiring standard prices of accessories calculated in advance according to statistics; obtaining second early warning information according to the standard price of the accessories and the current damage-assessment price; and obtaining target early warning information of the loss assessment price according to the first early warning information and the second early warning information.

In one embodiment, the processor, when executing the computer program, further performs the steps of: acquiring historical claim settlement data, wherein the historical claim settlement data comprises historical damage assessment data and historical damage assessment prices; performing feature vectorization on the historical damage assessment data to obtain a sample feature vector; and training according to the sample feature vector and the historical damage assessment price to obtain an early warning model.

In one embodiment, before the performing, by the processor when executing the computer program, feature vectorization on the historical damage data to obtain a sample feature vector, the method further includes: dividing historical damage assessment data into at least one class, and selecting a current sample center in each class; calculating the current distance between the historical damage assessment data in each class and the current sample center; calculating to obtain a new sample center of each type according to the first statistical value of the current distance; judging the difference value between the current sample center and the new sample center; when the difference value is smaller than the preset threshold value, taking the new sample center as a target sample center, otherwise, taking the new sample center as a current sample center, and continuously calculating the current distance between the historical damage assessment data in each class and the current sample center until the difference value is smaller than the preset threshold value; and carrying out standardization processing on the historical damage assessment data according to the target sample center.

In one embodiment, the normalization of historical damage data from a target sample center performed by the processor when executing the computer program comprises: calculating a first distance from the historical damage assessment data to the center of the target sample; calculating a second statistical value of the first distance to obtain a reference distance; and carrying out standardization processing on the historical damage assessment data according to the first distance and the reference distance.

In one embodiment, the inputting, by the processor, the current damage assessment data and the current damage assessment price into a pre-trained early warning model for processing when the processor executes the computer program to obtain first early warning information of the current damage assessment price includes: inputting the current damage assessment data into an early warning model obtained by pre-training to obtain a predicted price; when the predicted price is larger than or equal to the current loss assessment price, generating first early warning information that the current loss assessment price is normal; and when the predicted price is smaller than the current loss assessment price, generating first early warning information of abnormal current loss assessment price.

In one embodiment, the processor, when executing the computer program, further performs the steps of: and calculating a third statistical value of the historical damage assessment data in each class, and taking the third statistical value as the standard price of the accessories.

In one embodiment, the deriving of the second warning information according to the standard price of the part and the current damage price, which is realized when the processor executes the computer program, comprises: calculating discrete coefficients of the historical damage assessment data in each class, and calculating to obtain a reference threshold according to the discrete coefficients; and obtaining second early warning information according to the reference threshold and the current loss assessment price.

In one embodiment, the calculation of the reference threshold from discrete coefficients, as performed by the processor when executing the computer program, comprises: when the discrete coefficient is smaller than or equal to the threshold ratio, calculating to obtain the upper limit and the lower limit of a reference threshold according to the median and the standard deviation of the historical damage assessment data; and when the discrete coefficient is larger than the threshold ratio, calculating to obtain the upper limit and the lower limit of the reference threshold according to the average value, the median and the threshold ratio of the historical damage assessment data.

In one embodiment, the obtaining of the target warning information of the damage-rated price according to the first warning information and the second warning information, which is implemented when the processor executes the computer program, includes: when at least one of the first early warning information and the second early warning information is that the current loss assessment price is abnormal, outputting target early warning information with abnormal loss assessment price; and when the first early warning information and the second early warning information are not abnormal in the current loss assessment price, outputting target early warning information with normal loss assessment price.

In one embodiment, the outputting of the target warning information of the abnormal current damage price when the processor executes the computer program comprises: receiving an adjustment instruction aiming at the current loss assessment price; and adjusting the current damage assessment price according to the adjusting instruction, wherein the adjusted current damage assessment price is used for updating at least one of historical claim settlement data, the early warning model and the standard price of the accessory.

In one embodiment, a computer-readable storage medium is provided, having a computer program stored thereon, which when executed by a processor, performs the steps of: obtaining damage assessment information, wherein the damage assessment information comprises a current damage assessment price and current damage assessment data; inputting the current damage assessment data and the current damage assessment price into an early warning model obtained by pre-training for processing so as to obtain first early warning information of the current damage assessment price; acquiring standard prices of accessories calculated in advance according to statistics; obtaining second early warning information according to the standard price of the accessories and the current damage-assessment price; and obtaining target early warning information of the loss assessment price according to the first early warning information and the second early warning information.

In one embodiment, the computer program when executed by the processor further performs the steps of: acquiring historical claim settlement data, wherein the historical claim settlement data comprises historical damage assessment data and historical damage assessment prices; performing feature vectorization on the historical damage assessment data to obtain a sample feature vector; and training according to the sample feature vector and the historical damage assessment price to obtain an early warning model.

In one embodiment, before the computer program is executed by a processor to perform feature vectorization on the historical damage data to obtain the sample feature vector, the method further includes: dividing historical damage assessment data into at least one class, and selecting a current sample center in each class; calculating the current distance between the historical damage assessment data in each class and the current sample center; calculating to obtain a new sample center of each type according to the first statistical value of the current distance; judging a difference value between the current sample center and the new sample center; when the difference value is smaller than the preset threshold value, taking the new sample center as a target sample center, otherwise, taking the new sample center as a current sample center, and continuously calculating the current distance between the historical damage assessment data in each class and the current sample center until the difference value is smaller than the preset threshold value; and carrying out standardization processing on the historical damage assessment data according to the target sample center.

In one embodiment, the normalization of historical damage data from a target sample center performed by a computer program when executed by a processor comprises: calculating a first distance from the historical damage assessment data to the center of the target sample; calculating a second statistical value of the first distance to obtain a reference distance; and carrying out standardization processing on the historical damage assessment data according to the first distance and the reference distance.

In one embodiment, the inputting of the current damage assessment data and the current damage assessment price into a pre-trained early warning model for processing, when the computer program is executed by the processor, to obtain first early warning information of the current damage assessment price includes: inputting the current damage assessment data into an early warning model obtained by pre-training to obtain a predicted price; when the predicted price is larger than or equal to the current loss assessment price, generating first early warning information that the current loss assessment price is normal; and when the predicted price is smaller than the current damage assessment price, generating first early warning information of abnormal current damage assessment price.

In one embodiment, the computer program when executed by the processor further performs the steps of: and calculating a third statistical value of the historical damage assessment data in each class, and taking the third statistical value as the standard price of the accessories.

In one embodiment, the deriving of the second warning information from the standard price of the part and the current damage price when the computer program is executed by the processor comprises: calculating discrete coefficients of the historical damage assessment data in each class, and calculating to obtain a reference threshold according to the discrete coefficients; and obtaining second early warning information according to the reference threshold and the current loss assessment price.

In one embodiment, the computer program when executed by a processor implements calculating the reference threshold from discrete coefficients, comprising: when the discrete coefficient is smaller than or equal to the threshold ratio, calculating to obtain the upper limit and the lower limit of a reference threshold according to the median and the standard deviation of the historical damage assessment data; and when the discrete coefficient is larger than the threshold ratio, calculating to obtain the upper limit and the lower limit of the reference threshold according to the average value, the median and the threshold ratio of the historical damage assessment data.

In one embodiment, the obtaining target warning information of the loss price according to the first warning information and the second warning information, which is realized when the computer program is executed by the processor, comprises: when at least one of the first early warning information and the second early warning information is that the current loss assessment price is abnormal, outputting target early warning information with abnormal loss assessment price; and when the first early warning information and the second early warning information are not abnormal in the current loss assessment price, outputting target early warning information with normal loss assessment price.

In one embodiment, the computer program when executed by the processor, after outputting target warning information for a current damage price anomaly, comprises: receiving an adjustment instruction aiming at the current loss assessment price; and adjusting the current damage assessment price according to the adjusting instruction, wherein the adjusted current damage assessment price is used for updating at least one of historical claim settlement data, the early warning model and the standard price of the accessory.

In one embodiment, a computer program product is provided, comprising a computer program which, when executed by a processor, performs the steps of: obtaining damage assessment information, wherein the damage assessment information comprises a current damage assessment price and current damage assessment data; inputting the current damage assessment data and the current damage assessment price into an early warning model obtained by pre-training for processing so as to obtain first early warning information of the current damage assessment price; acquiring standard prices of accessories calculated in advance according to statistics; obtaining second early warning information according to the standard price of the accessories and the current damage-assessment price; and obtaining target early warning information of the loss assessment price according to the first early warning information and the second early warning information.

In one embodiment, before the computer program is executed by a processor to perform feature vectorization on the historical damage data to obtain the sample feature vector, the method further includes: dividing historical damage assessment data into at least one class, and selecting a current sample center in each class; calculating the current distance between the historical damage assessment data in each class and the current sample center; calculating to obtain a new sample center of each type according to the first statistical value of the current distance; judging the difference value between the current sample center and the new sample center; when the difference value is smaller than the preset threshold value, taking the new sample center as a target sample center, otherwise, taking the new sample center as a current sample center, and continuously calculating the current distance between the historical damage assessment data in each class and the current sample center until the difference value is smaller than the preset threshold value; and carrying out standardization processing on the historical damage assessment data according to the target sample center.

In one embodiment, the computer program when executed by a processor performs a normalization process on historical damage data according to a target sample center, comprising: calculating a first distance from the historical damage assessment data to the center of the target sample; calculating a second statistical value of the first distance to obtain a reference distance; and carrying out standardization processing on the historical damage assessment data according to the first distance and the reference distance.

In one embodiment, the inputting of the current damage assessment data and the current damage assessment price into a pre-trained early warning model for processing, when the computer program is executed by the processor, to obtain first early warning information of the current damage assessment price includes: inputting the current damage assessment data into an early warning model obtained by pre-training to obtain a predicted price; when the predicted price is larger than or equal to the current loss assessment price, generating first early warning information that the current loss assessment price is normal; and when the predicted price is smaller than the current loss assessment price, generating first early warning information of abnormal current loss assessment price.

In one embodiment, the obtaining target early warning information of the damage-assessment price according to the first early warning information and the second early warning information, which is realized when the computer program is executed by the processor, comprises: when at least one of the first early warning information and the second early warning information is that the current loss assessment price is abnormal, outputting target early warning information with abnormal loss assessment price; and when the first early warning information and the second early warning information are not abnormal in the current loss assessment price, outputting target early warning information with normal loss assessment price.

In one embodiment, the computer program when executed by the processor, after outputting target warning information for the current damage-price anomaly, comprises: receiving an adjustment instruction aiming at the current loss assessment price; and adjusting the current damage assessment price according to the adjusting instruction, wherein the adjusted current damage assessment price is used for updating at least one of historical claim settlement data, the early warning model and the standard price of the accessory.

It should be noted that, the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, presented data, etc.) referred to in the present application are information and data authorized by the user or sufficiently authorized by each party.

It will be understood by those skilled in the art that all or part of the processes of the methods of the embodiments described above can be implemented by hardware instructions of a computer program, which can be stored in a non-volatile computer-readable storage medium, and when executed, can include the processes of the embodiments of the methods described above. Any reference to memory, database, or other medium used in the embodiments provided herein may include at least one of non-volatile and volatile memory. The nonvolatile Memory may include Read-Only Memory (ROM), magnetic tape, floppy disk, flash Memory, optical Memory, high-density embedded nonvolatile Memory, resistive Random Access Memory (ReRAM), Magnetic Random Access Memory (MRAM), Ferroelectric Random Access Memory (FRAM), Phase Change Memory (PCM), graphene Memory, and the like. Volatile Memory can include Random Access Memory (RAM), external cache Memory, and the like. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM), for example. The databases referred to in various embodiments provided herein may include at least one of relational and non-relational databases. The non-relational database may include, but is not limited to, a block chain based distributed database, and the like. The processors referred to in the embodiments provided herein may be general purpose processors, central processing units, graphics processors, digital signal processors, programmable logic devices, quantum computing based data processing logic devices, etc., without limitation.

The technical features of the above embodiments can be arbitrarily combined, and for the sake of brevity, all possible combinations of the technical features in the above embodiments are not described, but should be considered as the scope of the present specification as long as there is no contradiction between the combinations of the technical features.

The above-mentioned embodiments only express several embodiments of the present application, and the description thereof is more specific and detailed, but not construed as limiting the scope of the present application. It should be noted that, for a person skilled in the art, several variations and modifications can be made without departing from the concept of the present application, which falls within the scope of protection of the present application. Therefore, the protection scope of the present application shall be subject to the appended claims.

Claims

1. A loss assessment price early warning method is characterized by comprising the following steps:

2. The method of claim 1, further comprising:

3. The method of claim 2, wherein before performing feature vectorization on the historical impairment data to obtain a sample feature vector, the method further comprises:

when the difference is smaller than a preset threshold value, taking the new sample center as a target sample center, otherwise, taking the new sample center as a current sample center, and continuously calculating the current distance between the historical damage assessment data in each class and the current sample center until the difference is smaller than the preset threshold value;

4. The method of claim 3, wherein the normalizing the historical impairment data according to the target sample center comprises:

5. The method according to any one of claims 1 to 4, wherein the inputting the current damage assessment data and the current damage assessment price into a pre-trained early warning model for processing to obtain first early warning information of the current damage assessment price comprises:

and when the predicted price is smaller than the current damage assessment price, generating first early warning information that the current damage assessment price is abnormal.

6. The method of claim 3, further comprising:

7. The method of claim 6, wherein the deriving second warning information according to the standard price of the part and the current damage-fixing price comprises:

8. The method of claim 7, wherein said calculating a reference threshold from said discrete coefficients comprises:

9. The method of claim 1, wherein obtaining target early warning information of the loss assessment price according to the first early warning information and the second early warning information comprises:

10. The method of claim 9, wherein after outputting the target warning information of abnormal current loss assessment price, the method comprises:

receiving an adjustment instruction for the current damage-fixing price;

and adjusting the current damage assessment price according to the adjusting instruction, wherein the adjusted current damage assessment price is used for updating at least one of historical claim settlement data, the early warning model and the standard price of the accessories.

11. A loss assessment price early warning device, characterized in that the device includes:

and the target early warning module is used for obtaining the target early warning information of the loss assessment price according to the first early warning information and the second early warning information.

12. A computer device comprising a memory and a processor, the memory storing a computer program, characterized in that the processor realizes the steps of the method of any one of claims 1 to 10 when executing the computer program.

13. A computer-readable storage medium, on which a computer program is stored, which, when being executed by a processor, carries out the steps of the method of any one of claims 1 to 10.

14. A computer program product comprising a computer program, characterized in that the computer program realizes the steps of the method of any one of claims 1 to 10 when executed by a processor.