A combined approach of formal concept analysis and text mining for concept based document clustering

Abstract

Nowadays, the demand of conceptual document clustering is becoming increase to manage various types of vast amount of information published on the World Wide Web. In this paper, we use formal concept analysis (FCA) method for clustering documents according to their formal contexts. Concept hierarchy of documents is built using the formal concepts of the documents in the document corpus. We use tf.idf (term frequency /spl times/ inverse document frequency) term weighting model to reduce less useful concepts from these formal concepts and the association and correlation mining techniques to analyze the relationship of terms in the document corpus.