Presentation information

General Session

General Session » [GS] J-13 AI application

[1P2-J-13] AI application: finance and economics

Tue. Jun 4, 2019 1:20 PM - 3:00 PM Room P (Front-left room of 1F Exhibition hall)

Chair:Yasuo Tabei Reviewer:Tomoya Yoshikawa

2:40 PM - 3:00 PM

[1P2-J-13-05] Automatic Summarization of Analyst Reports Based on Causal Relationships from News Articles

〇WATARU TAKAMINE1, Kiyoshi Izumi1, Yasunori Sakaji1, Hiroyasu Matsushima1, Takashi Shimada1, Yasuhiro Shimizu2 (1. University of Tokyo, 2. Nomura Securities Co. ,Ltd.)

Keywords:Text Mining, Extraction Causal Relationship, Representation Similarity

In this paper, we focused on the causal relationships in both of news articles and analyst reports. We proposed

a novel approach for summarizing analyst reports automatically based on the causal relationships extracted from

both text data. As a rst step toward summarization of analyst reports adequately, we analyzed the validity of

the method in extracting causal relationships which can be evaluated from the analyst reports. As a result, the

proposed method could extract basis information of analyst's opinions from analyst reports with some accuracy,

and we could conrm the styles of analysts in expression of opinions and bases.