Personalized Search Engine Based on Granular Computing

A binary system in information granules is introduced in this article. Based on this system, a search engine of web page abstract information is constructed, where the search results are processed by granular reasoning. Firstly, definitions for information granule and preprocess, such as accord degree, granular matrix, granular relation matrix, constrained relation matrix, association degree and so on, have been represented Secondly, with these definitions, a personalized meta-info search engine experimental system is developed. Finally, the experimental results show that the search results returned by this system match user's expectation effectively.