Affordable Access

Term Extraction based on the Forward and Backward Connectivities of Candidates

Authors
Publisher
情報処理学会
Publication Date

Abstract

In composite term extraction problems, it is important to extract candidates of relatively low occurrences in the corpora, with enough precision. In previous works, we have developed a method which can extract term candidates of low occurrences, using the revised classification of Japanese morphemes. In this paper, we propose a improved method considering forward and backward connective relations of candidates. Using the method, composite term candidates of less occurrences can be extracted with high precision.

There are no comments yet on this publication. Be the first to share your thoughts.