Research Of Abnormal Pattern Recognition Based On Data Intensive Methods

With the development of economy and society,urban vehicles increased dramatically,the traditional transport system appears more and more problems,traffic congestion,air pollution,traffic accidents became the focus of attention.At the same time,people travel quality requirements are also rising.In order to meet these challenges transportation,intelligent transportation systems have come into being,and in recent years has been rapid development.Intelligent transportation system through a combination of information technology and communication technology,a series of programs to solve the current traffic problems,and achieved some results,to ease traffic congestion and improve the quality of travel and other aspects have shown a tremendous value.In Intelligent Transportation Systems,one of the hot issues of the current study is the analysis of abnormal traffic.However,the conventional analysis of abnormal traffic problems exist,for example,many anomalies are monitored by manual inspection,not only can not make a scientific analysis and study of the traffic data,but also has a low accuracy rate,delay of the disadvantages.Therefore,the researchers are actively looking for new exception data analysis methods used to replace traditional methods of the past.In this paper,the existing data for learning algorithm analysis,combined with the characteristics of urban traffic flow,traffic data for a set of methods and processes abnormal analysis and K-means method to determine the starting point of the research data through.This method is characterized by analyzing the vehicle data,combined with real data verification,original,fuzzy vehicle was tagged data processing.Image processing also proposed model for the model to make adaptive changes characteristic data for different regions.And ultimately achieve the purpose of high accuracy traffic data clustering.This article is also based on data tagging after abnormal established classification model,aimed at the practical application of real-time traffic data to judge the possibility of an abnormality.This paper also presents a linear regression model based on the abnormality determination.Meanwhile,for data traffic characteristics,in order to improve the classification accuracy of the model,presented arguments to the collinearity,the traffic data set into suitable linear regression analysis of the data set,and ultimately improve the accuracy of the linear regression model in the study of traffic.Through the analysis of experimental results,methods and processes presented in this thesis to a certain extent,the problem effectively solve the anomaly of traffic data to help researchers raw traffic data initial recognition,and real-time traffic data classification and recognition processing.