Research on Blog Similarity Based on Ontology

Abstract

A new method to compute the similarity of two blog posts is proposed in this paper. This method mainly has two parts including keywords extraction and semantic similarity measurement. During keywords extraction part, the method utilizes particular post features to extract keywords from one blog post with the aim to improve the correlation rate. In order to compute the similarity of any two blog posts more effectively, semantic similarity measurement part make use of personal ontology that denotes single post, the aim is to transform the blog post similarity into the personal ontology similarity. This method has already been applied in an original retrieval system which supports searching not only by several keywords, but also by a blog post (in the type of URL). The experimental results demonstrate that the proposed method is effective.