结合社区结构的博客检索系统设计与实现-计算机科学与技术专业论文.docxVIP

Ab Abstract - - II - Abstract In recent years, with the rapid development of web2.0 era, blog is also explosive growing as an important Internet service. However, it also brings the problem of how to find the useful information from the massive amount of resources lying both in blog users and blog papers. To solve this problem, blog retrieval system which can provide the retrieval service in blogosphere is created. Through analysis, we find the problem of the main blog retrieval systems is that they are still not jumping out of the web retrieval thinking, in which both the demands of the users and some features of the blog resources are not taken into account. With these deficiencies in the current blog search service, this paper combined with community structural characteristics to try a count of new exploration. the main content can be summed up in the following areas: Firstly, Found on the analysis of existing methods of community detect, we proposes a method to find the potential community in the blogosphere based on some unique structural characters in blog in this paper. By extracting and filtering the blog label information, we get some directing information to find the topics which attract the blog author. Using these directing information and clustering methods, we completed the task of finding community. Our experiment result shows that this method can efficiently improve the accuracy of finding community in blog area. Secondly, based on the analysis and mining of latent community characters in the blogosphere, this paper proposes a blog author ranking method which combines the demands of blog retrieval system users with the structure of blogosphere community. This method takes into account the query relativity, the author topic tendency and the query topic tendency. This method is modeled and calculated with the result of latent community division. This paper’s method for blog authors ranking has distinct difference with the classical page ranking method