Blog Hot Post Discovery And Author Reputation Degree Research

At present, there are many researches about Topic Detection and Tracking, but these analyses aim at news events mainly. With the development of Web2.0, the using ways about Internet have changed a lot for people, namely, people's participation has increased significantly, so many application systems based on Web2.0 come into being. Blog is one of the main representatives. At present, Blog is a very important media on Internet, which can provide users with an easy way to express their own opinions and do comments on others' opinions, and comments which news events do not have are a main characteristic in BlogSphere. The freedom, arbitrariness and divergence characteristics in Blog field make it increasingly become the main place of generation and dissemination of public opinion. So, the importance on the research to Blog posts and Bloggers increasingly becomes prominent.Traditional Blog hot post discovery is obtained by the posts ranking of Blog web site, but this kind of ranking is based on the number of comments merely. The posts and comments' contents are not considered. Because of the high flexibility of comments, many noise comments appear inevitably, such as advertisements, hyperlinks and so on. In this paper, the main idea is that the number of comments to posts is only one of the factors to influence hot degree of posts, one post has more comments, and it has bigger possibility of being a hot post. On the premise of the number of comments, this paper analyzes the text contents of posts and comments, calculates the relation degree of contents between posts and comments firstly, and then does emotion analysis to the comments of lower relation degree, after this step, noise comments can be filtered. For left related comments, emotion analysis is used to compute the hot degree of posts, and then hot posts are discovered.Based on hot posts discovery, this paper does further analysis about the emotion of related comments for the Blog of higher hot degree posts. In this step, support degree of posts can be obtained by computing emotion of comments. Finally blogger reputation degree can be computed by support degree of posts obtained above. Experiments show that the approach in this paper can discover Blog hot posts and judge the reputation of their authors effectively. The method in this paper can help the management of the network's Blog space and network public opinion supervision.