Method Of Risk Level Discrimination For Operating Vehicle Drivers Based On Data Mining In Simulation Situation

As an important medium for road transportation,operating vehicles also have serious safety problems while achieving efficient and convenient transportation services.Among many factors,the driver of operating vehicles,as the leader of road transportation,plays a key role in the transportation safety.Therefore,effective risk assessment should be carried out for drivers of operating vehicles to screen out drivers with higher risks and supervise them in a targeted way.It is of great practical significance to prevent road transport accidents and guarantee the people’s life and property safety.Based on the driving data obtained by the simulation scenario,the data mining technology is used to establish an index screening model based on gray rough set,and then the risk level of driving is determined by using the grey clustering method based on the combination weight.Furthermore,a risk level discrimination model for operating vehicle drivers is constructed based on BP neural network.Therefore,based on the above theory and method,this paper realizes the effective identification of the risk level of drivers.The main work of this paper is as follows:(1)Taking the driving behavior of the operating vehicle driver as the research object,collecting the driving behavior data of 70 operating vehicle drivers with the help of the driving simulation experiment platform.This paper analyzes in detail the influence of driving behavior of drivers of operating vehicles on transportation safety,vehicle loss and other factors,and selects 16 driving behavior indicators preliminarily.(2)A method of index selection based on the combination of grey clustering and rough set attribute reduction theory is proposed to realize the selection of risk evaluation indexes for drivers of operating vehicles.It solves the problem that the traditional method can’t screen the indicators under the condition of uncertainty and incomplete data,and finally reduces 16 indicators to 9.(3)The driver’s risk level is divided into five grey categories: very low,low,general,high and very high.Based on the moment estimation theory,entropy weight method and order relation analysis method are combined to determine the weight of the evaluation index,and then grey clustering risk evaluation model based on combination weight is constructed by combining grey whitening weight function clustering.Finally,the risk assessment level of 70 drivers of operating vehicles is obtained.(4)Based on the internal learning principle of BP neural network,the initial network parameters are set up,and the driver risk level discrimination model is established,which takes the driver’s characteristic index as the input sample and the risk level as the output sample.After training and verification,it realizes the accurate identification of the risk level of the drivers of the operating vehicles.