Analyzing Public Opinion About Suzhou Traffic Based On Text Mining

Traffic public opinion means in certain social space,the public as the main part give opinions?comments or attitude on traffic environment?transportation or the trends of some traffic incidents.It is a total behavior of giving attitude?opinions or sentiments on all kinds of traffic phenomena,and it is also a comprehensive reflection of public opinion in transportation.With the rapid development of multimedia technology,more and more people take part in delivering traffic public opinion.Excavating information contained in traffic public opinion comprehensively is significant for controlling city traffic problems and their changes macroscopically and formulating a comprehensive urban transportation solution.Compared with traditional structured data generated in transportation,such as traffic flow and velocity,traffic public opinion is mainly unstructured data in the form of text?image or audio and video.It is possible to understand their value using technical methods such as intelligent classification?knowledge mining and so on.There are few studies on traffic public opinion at home and abroad at present,and the analytical system has not been established yet.This article means to explore how to build a system of analyzing traffic public opinion.In this study,with regard to traffic textual data elicited from internet forum,telephone hotline and traffic radio,different kinds of text mining methods are used for analyzing urban traffic public opinion.After preprocessing of the textual data,topical subjects of traffic public opinion are automatically classified by using the SVM model;characteristics and differences between data sources are revealed by Correspondence Analysis;traffic phenomena hidden in key-words are excavated by Association rules based on Apriori algorithm;traffic problems reflected from key-words and their changes over time are further clarified by Co-occurrence network analysis.The elements that influence people complaint were found through empirical analysis,and the impacts brought by road construction?reconstruction are the most obvious,and according to the analysis on different phrase of specific road reconstruction,the implementation of new schemes after road reconstruction is the most controversial issue.Meanwhile the association between key words of different type of traffic public opinion and some regularities of traffic congestion in the city have also been found out through this study.The text mining process proposed by this article make contributions to studying traffic public opinion data further.And the results from this article can provide the reference for the traffic management department on how to improve people's transport satisfaction and create a better traffic environment.