| With the rapid development and popularization of network technology, people’s lives become increasingly dependent on the Internet, and the Internet has gradually become a major platform for people to obtain information and exchange information, due to the convenience and freedom of the network messaging features, positive and negative information on the Internet spread very quickly become a very good response to the social moment of public opinion in a major media network, these negative public opinion if not properly controlled society may have a huge impact, while some positive public opinion, access to information to be able to make the right decisions for government agencies to provide a supporting role. Particularly evident in the role of the network public opinion in a network environment as a network of more than90%of the college complex network environment can not only education in schools and day-to-day management of huge benefit, but also for some of the negative information dissemination provide a good platform, so the university network monitoring public opinion has become urgent and important. Forum as the school’s most active network platform has become the most important way of public opinion spread. This thesis focus on information capture and analysis of public opinion for college forum. In view of the mass characteristics of the network public opinion, and then artificial means to collect and analyze the cost becomes high and efficiency is not high. Research and Realization of this university forum of public opinion capture and analysis system. Good decision-making role in support of the system for the university management.The in University Forum characteristics, this article focuses on the following key issues discussed and R&D:multi-threaded focused web crawler, high efficiency of Chinese word segmentation, text orientation analysis. Firstly collection techniques on the Web page that the Web crawler, and the principle of Colleges and Universities Forum features fully integrated, an orientation of public opinion on campus BBS data collection and extraction system. And focus for the forum network reptiles, joined the multi-threading technology, greatly improving the work efficiency of the Web crawler. To make appropriate changes in the use of the the Pangu word method based on Chinese word, word of concern highlighted function. Before segmentation, the method will take three Dictionary advance loaded into memory Dictionary, and using a hash function to speed up the reading efficiency of the Dictionary, and thus very efficient segmentation. Based on the study of the existing Chinese opinion analysis, this paper presents the a tendentious analysis method for the BBS topic posts and thread. |
