Abstract

We present discussion mining as a preliminary study of knowledgediscovery from discussion content of offline meetings. Our systemgenerates minutes for such meetings semi-automatically and links themwith audio-visual data of discussion scenes. Then, not only retrieval ofthe discussion content, but also we are pursuing the method of searchingfor a similar discussion to an ongoing discussion from the past ones,and the method of generation of an answer to a certain question based onthe accumulated discussion content. In terms of mailing lists and onlinediscussion systems such as bulletin board systems, various studies havebeen done. However, what we think is greatly different from the previousworks is that ours includes face-to-face offline meetings. We analyzemeetings from diversified perspectives using audio and visualinformation. We also developed a tool for semantic annotation ondiscussion content. We consider this research not just data mining but akind of real-world human activity mining.