Construction of Association Word Knowledge Base

초록

Query expansion in knowledge based on information retrieval system requires knowledge base being considered semantic relations between words. We extract association word using Apriori algorithm. In that case, recall and accuracy become higher. This paper shows how to construct association word knowledge base with improved accuracy, including association words being considered semantic relations between words. Toward this end, web documents related to computer are classified into eight classes, and nouns are extracted from web document of each class. Association word is extracted from nouns through Apriori algorithm. In order to evaluate performance of Apriori algorithm suggested in this paper, Apriori algorithm has been compared with mutual information and Rocchio algorithm.