The Prague Stringology Club Workshop '98

Kazuaki Ando, Kimihiro Iwasaki, Masao Fuketa and Jun-ichi Aoe

A Fast Morphological Analysis Using the Extended AC Machine for Oriental Languages

Abstract:
This paper presents a fast morphological analysis for oriental languages by extending an Aho and Corasick's pattern matching machine. Our method is a simple and efficient algorithm to find all possible morphemes in an input sentence and in a single pass, and it stores the relations of grammatical connectivity of adjacent morphemes into the output functions. Therefore, the costs of checking connections between the adjacent morphemes can be reduced by using the connectivity relations. Furthermore, the construction method of the relations of grammatical connectivity is described. Finally, the proposed method is verified by a theoretical analysis, and an experimental estimation is supported by the computer simulation with a 100,267 words dictionary. From the simulation results, it turns out that the proposed method was 49.9% faster (CPU time) than the traditional trie approach. As for the number of candidates for checking connections, it was 25.5% less than that of the original morphological analysis.

Download article: Article in PostScript Article in PDF
 PostScript   PDF