The Safety Assessment For The Road Transportation Enterprise Based On Improved Back Propagation Neural Network

Road passenger transportation plays an important role in road transport,and has a direct impact on the safety of the residents travel and the stable development of transportation.At the same time of the volume of passenger transportation is increasing, the safety work of road passenger transportation enterprise is backward, they can not be coordinated development, so resulting the traffic accidents always occurred, hindering the harmonious development of the state and society.Safety assessment effectively improves the enterprise's safety management work, and the awareness of the staffs,so employees have the ability to investigate hidden danger of accident and responsd to emergency,enhancing the preparations for the accident prevention,and improveing the safety management of enterprises.Through the analysis of the development and safety of road transportation in our country, the transportation enterprises play an important role in the road traffic safety.Road passenger transportation occupies the main status in the entire transportation system, this thesis researches the safety management work of Road passenger transportation enterprise, according to the research on the basic theory of safety assessment. Analysis the main factors that affect the security of the enterprise, studing the two impacts from the enterprise's internal management and safety system, combined with the enterprise situation,constructing a reasonable evaluation index system, and designing the safety evaluation scale.Comparing the applicability of various evaluation methods,the analytic hierarchy process(AHP) suitable for evaluation problem that indexes are less, so this method was used to study the driver's accident causes.Principal component analysis method uses less indexes instead of the original complex and different indexes, and improved BP neural network has high accuracy, small workload characteristics, the combination of the two methods can effectively evaluate the safety in enterprise.Choose the some example of enterprise, getting scores through the security check scale ofenterprises, principal components was extracted by principal component analysis method,inputting the processing data to improved BP neural network.Evaluation results show that the model error and the expected value between 0.2% and 1%, the error is small, indicating that this evaluation method is practical, and has the characteristics that the workload is small,can evaluate the safety of the enterprise.