Research And System Development Of Highway Traffic Public Opinion Monitoring Technology

The Internet has provided a convenient way for the distribution of public opinion content,thanks to the rapid growth of Internet and communication technologies.More and more information will be accessible to the public via the Internet.Public opinion information detection and analysis has transited from artificial analysis to a series of text sentiment analysis algorithms.Public opinion information analysis could be a way to find the public opinion orientation of the whole society,and help decision-makers make corresponding response for emergencies and make adjustments if there is a false direction of public opinion,which is beneficial to the stable development of our society.In view of the problems above,a set of public opinion monitoring and analysis systems is redesigned and developed in this paper which includes the following jobs:(1)Public opinion data collection.This paper uses web crawler technology to collect public opinion information data from Jiangsu highway network.Through the research on the interaction between traditional crawler’s downloader and parser,web page parsing and other aspects,the problems of incomplete and timeliness of crawling public opinion data existing in the traditional crawler are resolved.(2)The system design and implementation and related technologies such as feature selection,Chinese word segmentation,classification algorithm,text representation are analyzed and introduced.(3)Based on the above crawled data sets,this paper uses the long-term memory network(Long short-term memory,LSTM),convolutional neural network(Convolutional Neural Network,CNN)and support vector machine(Support Vector Machine,SVM)to design and train the above three methods.And a series of evaluation criteria such as accuracy,precision and recall rate are used as evaluation indexes.Finally,the results of these three algorithms are better applied to the public opinion monitoring and analysis system.(4)Finally,a set of public opinion monitoring and analysis system is written by java language,which includes user login,public opinion retrieval,public opinion analysis and other functions.Finally,the long short term memory(LSTM)algorithm is applied to the public opinion monitoring and analysis system.The results show that the long short term network is used in the test and analysis system Compared with the other two algorithms,memory(LSTM)algorithm can obtain the best performance results,and the accuracy,precision and recall rate reach 96.1%,84.2% and 88.9%respectively.When the threshold value is 0.7,the overall performance is the best,the AUC value can reach 0.904,and the indicators are also improved.