A Prediction Of Urban Short-time Traffic Flows Based On Bayes Classification And K Neighbor Method

With the rapid development of economy society,urban road traffic congestion has become the top issue facing the word cities,how to reasonable management and control of urban road traffic,is the traffic management department and transportation scholars have problem in thinking and research.In this background,the intelligent traffic management system by traffic management department and transportation more and more scholars' attention.The system of some subsystems such as traffic guidance system,traffic control system can effectively alleviate the problem of urban road traffic congestion.Urban road traffic flow information of the real-time and accurate is the foundation of the intelligent traffic management system give full play to the role.Of urban road short-time traffic flow prediction based on the predecessors' research,based on the summary and analysis about the basic theory of traffic flow forecast,put forward in this paper,the research methods and improving direction,and the feasibility of the improved prediction method is verified by the simulation calculation,therefore,in this paper,the main work is as follows:(1)The first two chapters of this article is mainly on the research literature at home and abroad and referring to do after the analysis and summary.First of all,the first chapter describes the research background and significance,research status both at home and abroad as well as the content and structure of the paper,etc.The second chapter is elaborated according to the prediction of traffic flow forecasting of the length of the interval classification and the classification,the main purposes of summary and analyzes of the main traffic flow prediction method and the advantages and disadvantages of each method,traffic flow data collection and pretreatment method is given.(2)The third and fourth chapters two of this article is to provide short-term traffic flow forecasting method adopted in this paper.In this paper,the prediction method is adopted by the parameter regression prediction method,in the third chapter first summarizes the non-parametric regression prediction principle,advantage and application in the field of short-time traffic flow prediction.And then introduces the non-parametric regression method and commonly used in the kernel function method and K neighbor method,on the basis of a K neighbor method as the forecast method of this paper.And then put forward the improvement in K neighbor prediction model of two directions,one improvement is the constant K value instead of on the basis of traffic flow state of different selection of K value,improve the second is to consider intersection,introduce related to upstream intersection traffic flow.Due to the variable value K,need to categorize traffic flow state,in the fourthchapter introduces the based on bayesian classification K neighbor short-time traffic flow prediction.And the forth chapter is mainly introduced the bayesian theory,and this paper USES the simple bayesian method,using naive bayesian network and introduces the state of historical data and predicted data classification method,and the use of bayesian classification in this paper,the prediction process.(3)In this paper,the simulation result of the fifth chapter,first validated in different state of traffic flow under the premise of different traffic flow status,you can pick and choose the corresponding the best K value,and on the premise of no traffic state classification,you can select the best K value.Secondly,this paper is verified in the third chapter,the feasibility of the mentioned two improvement direction.Then two kinds of improved at the same time use,and was the improved K neighbor method of compare forecast results and analysis,compare the error indicators,prove the improved method better prediction precision,error distribution is more good,can be used in urban road short-term traffic flow prediction.Finally,the study of this article made a summary,and points out the next research direction.