論 文Papers

CONFERENCE (DOMESTIC)

株価掲示板情報における煽り情報の検出

山下 達雄, 坪内 孝太

2015年度人工知能学会全国大会(第29回), 2015/5

Category:

自然言語処理 (Natural Language Processing) 機械学習 (Machine Learning)

Abstract:
A few users submit the agitate information in the stock bulletin board for control the stock price. For instance, agitator submit fake information or information from opposite stance to actual. These information deteriorates reliability of the bulletin board and be a strong noise in data analysis. This paper propose the way of agitation detection of stock bulletin board by analyzing users’ article. The simulation test with 2 years bulletin board data shows the proposed method can detect agitation accurately.
Download:

株価掲示板情報における煽り情報の検出(外部サイト/External Site Link)