Research And Implementation Of Key Technologies Of BLOG Search Engine Based On Ontology

With the rapid expansion of blog information, it is becoming more and more important to get blog information from the Internet effectively and precisely. Under this background, many researchers have paid increasing attention to the related technology on blog search engine.However, the existing blog search engines still have some problems. First, the engines do not meet the requirement of document level search in the way of inputting a post (blog articles) and returning relevant posts because they do not have a valid method on semantic similarity calculation of posts. Second, the results from the engines are inaccurate for the imprecise value of semantic similarity calculation. Last but not least, the answers relevant are not always highlighted and ordinal, which is due to the low-effect of result ranking algorithm.In order to solve these problems, this paper makes some research on semantic similarity calculation and result ranking algorithm, to make the search results become more precise, more ordinal and satisfy the user requirements.On the basis of previous studies on semantic similarity calculation, we propose a method computing the semantic similarity of posts based on Ontology named SSPO in this paper. The offered method not only means the similarity is semantic similarity, but also supports for computing the similarity of two posts by inputting posts directly. So we can realize the document search by using SSPO. In order to improve the accuracy of semantic similarity calculation, SSPO utilizes the features of posts to optimize the performance of the keyword extraction algorithm which will benefit the similarity accuracy.There are many result ranking algorithms like HITS, PageRank and so on, while we focus on the PageRank algorithm here because of its wide influence. PageRank also has some deficiency, like emphasis on the old page, neglect of professional site and migration of the theme. There are also many other improvements on PageRank, but these improvements can not conquer all of these shortages. So in this paper, we offer IPageRank, which associate the PageRank algorithm with the consideration of the semantic similarity of posts, to improve the original algorithm.This paper also constructs area ontology on Shanghai World Expo for Prototype System BSE., which is the application system for the SSPO and IPageRank. In the end, by simulating experiments, it is able to demonstrate the validity of SSPO and IPageRank algorithm.