Behavioral anatomy of asset price fluctuations based on large textual data.

2015 Fiscal Year Final Research Report

• Project/Area Number: 25285127
• Research Category: Grant-in-Aid for Scientific Research (B)
• Allocation Type: Partial Multi-year Fund
• Section: 一般
• Research Field: Management
• Research Institution: Kwansei Gakuin University
• Principal Investigator: OKADA Katsuhiko 関西学院大学, 経営戦略研究科, 教授 (90411793)
• Co-Investigator(Kenkyū-buntansha): HAMURO yukinobu 関西学院大学, 経営戦略研究科, 准教授 (90268235) ; KATOH naoki 関西学院大学, 理工学部, 教授 (40145826)
• Project Period (FY): 2013-04-01 – 2016-03-31
• Keywords: ビッグデータ / 人工知能 / テキストマイニング / 投資家心理 / ハーディング / 行動ファイナンス / アノマリー

Outline of Final Research Achievements

The purpose of this research is to identify the role of investor psychology in the stock market. For this purpose, we have collected a massive amount of news data as well as the social data. (ex. 5.7 million news paper articles. Five years of data in the stock market messages boards. Past one year data of the number of visitors of company web-site.) These large textual data are then organized in a way that we can identify to what company each news text belongs. Simultaneously, we have created a dictionary in order to determine whether the text is positive or negative. Based on these data base, we have developed a time varying "Sentiment index" for each one of the listed stocks.
We found stock price reactions to a fundamental shock such as earnings announcement surprise / securities analysts' rating revisions are subdued when the sentiment index is in line with the shock, whereas it becomes large when the sentiment is opposite of the surprise.

Free Research Field

行動ファイナンス、人工知能、