Research Of Key Technology On Internet Public Opinion Monitoring System

With the rapid development of Internet and mobile Internet, the Internet has become an important platform for informational communication and collision. Internet public opinion as an extension of social public opinion can fully and truly influence the various problems that exist in society. The study of Internet public opinion is important to understand and solve problems. This paper focuses on the technology and algorithms that related to monitoring Internet public opinion. The study also designs and implements a real-time environmental public opinion monitoring system. The main work of this article includes the following aspects:First, this dissertation introduces the current situation of the researches about Internet public opinion and environmental Internet public opinion monitoring and analysis system at home and abroad.Second, this dissertation does research on the web crawler. It designs a distributed web crawler framework by using asynchronous non-blocking I/O technology to achieve concurrency. According to different types of network public opinion data source designs different solutions. In this way, this framework can get data faster and more accurate.Third, this dissertation studies the key technology about text classification and does experiments on different algorithms. According to our experimental results this dissertation chooses SVM to classify environmental Internet public opinion.Fourth, after the research on the algorithms of construct tendentious views dictionary, this dissertation designs a term weight calculation method based on view propensity weights. Experimental results show that this method can improve the effect of judging text tendency greatly.Last, this dissertation designs and implements an environmental Internet public opinion monitoring system. This system can auto collecting and processing information and recognize predefined text sensitive topics related to environmental protection. It also can detect, identify tendentious and analysis the information. By showing the analyticalresult clearly, this system can let relate department easy to get the first time public opinion.