Yoo Chun Sik, Woo Sun Mi, Yoo Cheol Jung, Lee Jung deuk, Gwun Ou Bong, Kim Yong Sung

The Transactions of the Korea Information Processing Society (1994 ~ 2000), Vol. 5, No. 6, pp. 1552-1562, Jun. 1998
10.3745/KIPSTE.1998.5.6.1552,   PDF Download:

Abstract

Typical techniques for automatic indexing on Korean documents are having difficulty in linguistic processing such as morphological analysis because a part of speech is vague and the process of complex noun is not easy. Also, a separate stop-word list have to be developed upon each different subject and the size of a stop-word list is hug to control. To solve these problems, this theme proposes an automatic indexing technique which performs simple morphological analysis without analysis upon complex nouns and upon process of ambiguity of a part of speech. After constructing finite automata using simple nouns extracted through a morphological analysis, importance of words in text are figured out according to the term frequency of simple nouns and finite automata. Consequently, overhead upon ambiguity of complex nouns and a part of speech is reduced. This theme investigated performance of a proposed automatic indexing technique in comparison with indexing words extracted through a manual indexing process.

Statistics
Cite this article