Discovery Of Implied Communities For Blog Page Based On Topic

The methods about discovery of implied communities for Blog page based on topic bring the Blogs' applied research into a promising and challenging new domain. Along with its rapid development, Blog pages have been used as a new Internet media communication tools changing aspects of our culture , economic , politics or mainstream media for its own unique characteristics. Considering the Blogs' special features, common Blog search engine technologies can not satisfy the people's needs concerning certain topics. So We have to apply a combination of information retrieval, data mining, integrating information extraction and natural language processing techniches into the Blog content analysis and page link analysis is necessary to various Blog applications, such as analyzing the public opinions on certain product or policy, finding our interest, mining people's potential emotion about certain events . Although we also can analyze relation between Blogger's idea or Blogger's view and mainstream media or social culture according to Blogger's action and Blog entry's description. This paper studies topic-based Blog implied communities discovering in both theoretical and practical point of view. Topic-based Blog implied community discovering is a challenging domain along with the accelerating development and increasing influences of Blogs pages in the Internet . We began our research work from above perspectives and achieved the following achievements:In the thesis, we propose a SPC (Search Path Count) based Topic-based Blog implied community discovering algorithms. In the algorithms, we suggest a new method for scoring the relevance between a Blog entry and a topic based on the analysis to Blog page content and page link. Meanwhile, we introduce the graph theory to our algorithms to detect the Topic-based Blog implied community. The experiments show the new algorithms has improved the coverage rate and the interesting coefficient obviously.In our experiments, we first studied the relations between the Blog entries and the links and their influences to the detection of Topic-based Blog implied community. Then, concerning their difference in the specific community, we studied the relations between topics and Blog link feature and analyzed the effects of different topics to the activity and the significance of Blogs by extract top 10 from computing Eigenvector weights set. We applied SPC (Search Path Count) method to the Topic-based Blog implied community discovering in our Blog objected search engine system. We automatically crawled Blog entries from internet and computing various weights, then we displayed the retrieval results as topic based clusters. The experiments showed better user satisfactions and improved relevance of the retrieval results.