Accident vehicle determination method based on deep learning algorithm

Abstract

The invention provides an accident vehicle determination method based on a deep learning algorithm, and the method is characterized in that the method comprises a model training stage: importing sample data, each piece of sample data comprising a vehicle diagnosis report, a vehicle fault maintenance case and a maintenance file; using a random forest algorithm for constructing a model, substituting sample data for model training and evaluation, wherein the executed operation comprises the steps that fault information in a vehicle diagnosis report is extracted, and obtaining frequently-replaced parts corresponding to the fault information through mapping of the fault information and a vehicle fault maintenance case; according to the relationship between the parts and each system class in the vehicle, carrying out corresponding conversion to obtain the weight score of each system class; finally, obtaining the weight score of each sample classification relative to the fault information; and a probability analysis stage: extracting fault information in the vehicle diagnosis report, substituting the fault information into the trained model, and outputting the classification probability or / and decision result of each sample after model analysis.

Description

technical field [0001] The invention is used for the probability judgment of accident cars in the second-hand car market, and specifically relates to a method for judging accident cars based on a deep learning algorithm. Background technique [0002] With the support of national policies, the number of domestic vehicles and the rapid increase of buyers, the transaction volume of the second-hand car market is increasing day by day. The minefields that cannot be avoided will affect the driving experience at least, and endanger the safety of life and property at the worst. At present, accurate accident car judgments are based on maintenance files, but it is difficult to obtain maintenance records that are strictly controlled. cannot be judged. In addition, there are other judgments of accident vehicles that are not generated by maintenance files, and such judgments will have large judgment errors due to technical means. These are not beneficial for buyers to judge the condit...

AI Technical Summary

Problems solved by technology

At present, accurate accident car judgments are based on maintenance files, but it is difficult to obtain maintenance records that are strictly controlled. it is impossible to judge

In addition, there are other judgments of accident vehicles that are not generated by maintenance files, and such judgments will have large judgment errors due to technical means.

Images