Public Transportation Passenger Flow Prediction Based On Deep Learning Method Research

In recent years,China’s economy has been rapid development,urbanization speed significantly accelerated,the rapid increase in population in the city,resulting in the increasing demand for motor vehicles.How to build smart city,intelligent transportation system,is a valuable subject.Intelligent transportation system(Intelligent Transportation Systems,ITS)becomes the key way to solve traffic problems.City traffic flow includes not only the traffic in the city,the city should also include customer flow.Accurate prediction of urban traffic flow is a very important part of ITS,accurate grasp of urban traffic flow trends,for the construction of smart city is of great significance.In recent years,deep learning has become a hot topic in science and technology.Deep learning has become a branch of research in machine learning because of its strong learning ability.At present,depth learning is widely used in the industrial field to other fields,such as image recognition,Natural Language Processing,especially intelligent language recognition and many other high-tech fields.However,it is an innovative research direction to apply depth learning to deal with traffic flow problem.Through the prediction of urban traffic flow,can provide decision-making for the public travel,improve urban traffic efficiency,reduce congestion in the city,reduce emissions,improve air quality,have a positive mean.The focus of this paper is to predict the short-term passenger flow by depth learning.The main research contents are as follows:(1)This article will briefly introduce the current situation of public transport and the depth of the development process of learning.(2)The origin of the depth study,the research progress in recent years,as well as the classical model and the training algorithm in recent years.(3)Combined with the depth of the depth of the encoder model,according to the existing BRT in Shaoxing,China’s credit card records and outside weather,holiday factors,to predict the short-term BRT passenger flow.(4)Depth study,as a deep structure model,has excellent performance in dealing with data dimension reduction.The use of the classical model of the shallow layer model SVM fusion depth structure,processing traffic data processing.