Word Segmentation in Japanese for Construction of Scientific Japanese Corpus

摘要

Japanese word segmentation in the specialized field is the difficulty of building the Scientific Japanese Corpus. This paper deeply studied the basic problem of the word segmentation. They are ambiguity and words without login. In this paper we tried to use the ART network and immune principle to solve the problem of Japanese word segmentation, more than 90% of the words are correctly separated by our method.