会議名: 情報知識学会 第11回(2003年度)研究報告会
While search technologies have made marked progress in recent years, it is yet time consuming and laborious to read references output after searching. Particularly so, to Japanese reading English journals. From our experiences, we knew that sentences having numbers are more meaningful in an article. We came to an idea that if such number-containing sentences are extracted, it will be helpful for a reader to quickly grasp an outline of the article. This report is a result of our examination of the theory, and we developed a PC tool for extracting such number-containing sentences.