No abstract

No abstract

No abstract

Weblogs (blogs) becomes a very popular medium for exchanging information, opinions and experiences nowadays. However, since new blog pages are constantly issued, finding out helpful information from them becomes a tedious and time consuming work. This paper proposes a system for extracting knowledge hidden in blog pages in Chinese. Before extraction, blog pages are clustered into categories. Then for each category, the knowledge can be extracted based on domain ontologies. Using restrained natural language processing, user can query the KB and the helpful knowledge will be returned based on reasoning about the individuals. KEROB, a prototype of our system, is designed and implemented to fulfill the above functions. The experimental results indicate the superiority of our system.