Mining from literary texts

pattern discovery and similarity computation

Masayuki Takeda

Tomoko Fukuda

Ichiro Nanri

pp. 518-531

in: Setsuo Arikawa, Ayumi Shinohara (eds), Progress in discovery science, Berlin, Springer, 2002

Abstract

This paper surveys our recent studies of text mining from literary works, especially classical Japanese poems, Waka. We present methods for finding characteristic patterns in anthologies of Waka poems, as well as those for finding similar poem pairs. Our aim is to obtain good results that are of interest to Waka researchers, not just to develop efficient algorithms. We report successful results in finding patterns and similar poem pairs, some of which led to new discoveries.