| One big advantage of the development of the Internet is it brings people anytime, anywhere to participate in social activities, and it is convenient for them to make a comment for the interesting topic, with the development of social networking technique, tweeting has become a very convenient way for communication. The characteristics of the micro blog is convenient, fast and no threshold, a person who has large number of fans, their blogs can easily create enormous influence, so the celebrity, the ordinary people are among the micro blog, recording their daily lives, spreading their ideas.Micro blog gives us lots of convenient, but also produced a lot of problems, rumors and false information, fresh news no longer need to spread by traditional media, Micro-blog can quickly transfer messages, at the same time the cost is very low. In order to create high impact, many public relationship companies hire the online ghostwriters to hype. Therefore, it is very necessary to analysis of the spread of micro-blog and verify the truth of it. Different from simple text analysis, we need a method to propagation analysis and truth verify for micro blog.This paper firstly introduces the research background and the project done by predecessors, then introduces the methods and techniques used by the algorithm and the technology for the project, including sentiment analysis, machine learning, statistics, and engineering. And then put forward a method for distinguishing zombie users, a method for micro-blog opinions leader recognition, and then use the statistical analysis of the data transmission as features, use machine learning methods to train a model to make verification. Finally using the model,using engineering technology, based on the needs of users made a program for hot micro-blog analysis, data display, material selection for user to use. |
PDF Full Text Request