Research On Public Opinion Guidance Technology Based On Improved PageRank Algorithm

With rapidly developing of internet,the web is gradually involved in people's daily life, which leads to the booming of social network at the same time.As the virtual nature and anonymity of the network,the cyber citizen can express their thoughts on a variety of hot events freely. As more and more people are involved, it has become the public sentiment.The range and speed of network public opinion is far beyond the traditional public opinion, which reflects the views of the public directly and rapidly so that its influence is growing larger and larger. How to supervise and guide the network public sentiment has been a major challenge to leaders and local government officials.Faced with such complex data, if we use the human monitoring method to solve this problem, there will be many unreasonable factors coming up, so the information mining technology has become a best choice to solve the problem. At present, the domestic and foreign research on the technology of public opinion is relatively small, only focus on the technologies in emotion analysis of public opinion information, topic tracking and other. For this, this thesis presents a framework for guiding the public opinion. this framework includes the way from the monitoring and collection of the relevant information of public opinion to the construction of a corpus of public opinion to guide a set of functions.In the framework, there are two core issues need to be focus on:the reply corpus is hard to collect and Information publisher is difficult to choose. Aiming at these problems above, this paper presented a new algorithm to collect the corpus based on the idea of the PageRank and text sentiment analysis. To solve the problem of how to choose the publisher, this paper proposed an algorithm based on the emotional tendency of replies. Experimental results on the corpus of Tianya Forum show that the proposed method can achieve better performance.Finally, this paper carries out the implementation of the core modules and content based on the public opinion guidance framework.