An Analysis of Questions in a Q&A Site Resubmitted Based on Indications of Unclear Points of Original Questions

Proceedings of the 5th International Joint Conference on Natural Language Processing, pages 1428–1431, Chiang Mai, Thailand, November 8 – 13, 2011. c⃝2011 AFNLP An Analysis of Questions in a Q&A Site Resubmitted Based on Indications of Unclear Points of Original Questions Masahiro Kojima and Yasuhiko Watanabe and Yoshihiro Okada Ryukoku University, Seta, Otsu, Shiga, 520-2194, Japan t10m101@mail.ryukoku.ac.jp, {watanabe, okada}@rins.ryukoku.ac.jp Abstract In this study, we analyzed how answerers indicated unclear points in questions, and how questioners modified and resubmitted their questions based on indications of un- clear points. 1 Introduction In these days, many of us use question and answer (Q&A) sites where we share our problems and get solutions of them. For example, about 3.11 million questions were submitted to Yahoo! chiebukuro 1 from April/2004 to October/2005. Because of this large numbers of questions, questioners had better submit questions which give enough information to answerers. However, it is difficult to make good questions. For example, in Yahoo! chiebukuro, we often found unclear questions (e.g. Q1 in Figure 1) and their answers where answerers indicated un- clear points of the questions (e.g. A1 in Figure 1). (Q 1) I cannot access a web page which I could read yesterday. What should I do? (A 1) Show URL. In (A 1), the answerer pointed out that the ques- tioner did not describe important information to an- swer the question: URL. Unclear questions may de- crease chances of getting good answers. As a result, it is important to investigate supporting methods of making clear questions. One idea is to indicate un- clear points of questions, as the questioner of (A 1) did. In order to obtain helpful knowledge and de- velop a help system for making clear questions, it is important to analyze • how answerers indicated unclear points in questions, and • how questioners modified and resubmitted their questions based on indications of unclear points in their original questions. 1Yahoo Answers in Japan. http://chiebukuro.yahoo.co.jp Figure 1: An example of a resubmitted question based on the indication of unclear points in the orig- inal question. Our approach differs from previous analyses on Ya- hoo! Answers (Su et al., 2007) (Adamic et al., 2008). In this study, we used the data of Yahoo! chiebukuro for observation and examination. The data of Yahoo! chiebukuro was published by Ya- hoo! JAPAN via National Institute of Informatics in 2007 2. This data consists of about 3.11 million questions and 13.47 million answers which were posted on Yahoo! chiebukuro from April/2004 to October/2005. 2 Types of indication of unclear points in questions and modification of questions 2.1 Types of indication of unclear points in questions We observed answers submitted to PC category of Yahoo! chiebukuro and found the following five types of indication of unclear points in questions (Table 1). TYPE (A1-1) Answerers wanted detailed ac- counts of what questioners did. (Q 2) I got a warning from Symantec. How do I extend the software license. (A 2) Did you buy an extension key? 2http://research.nii.ac.jp/tdc/chiebukuro.html 1428 Table 1: …

Content truncated to 500 words