A Text Mining System DIREC: Discovering Relationships between Keywords by Filtering, Extracting and Clustering

摘要

This paper presents a text mining system DIREC that obtains the relationships between keywords. After gathering Web pages related to a query with search engines and filtering out the pages irrelevant to the query, DIREC extracts pre-specified keywords from the pages left after filtering, and clusters them. Each cluster includes keywords related to each other. The clustered keywords are shown through an Explorer-like graphical user interface. We discuss the experimental results applying the DIREC system to the Research and Education Report Database of the Faculty at Kyushu University.