Detecting Mandarin question sentences is both interesting and difficult. To tackle this new topic, our strategy is first to try to increase recall and then precision. To achieve higher recall, we not only review relevant linguistic literature but also re-examine relevant issues from a new statistical and corpus point of view, and discover more comprehensive and precise question-related words than before. Next we present our statistical approaches and procedure, and discuss our findings. We achieve good recall and modest precision in the preliminary study, and pioneer the computational study of indefinitives.
|頁（從 - 到）||466-478|
|期刊||Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)|
|出版狀態||Published - 1 十二月 2003|