Development of an information recommendation framework based on item operations that reflects human feelings about objects and things

2020 Fiscal Year Final Research Report

• Project/Area Number: 18K11551
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Review Section: Basic Section 62020:Web informatics and service informatics-related
• Research Institution: Kogakuin University
• Principal Investigator: Kitayama Daisuke 工学院大学, 情報学部(情報工学部), 准教授 (40589975)
• Project Period (FY): 2018-04-01 – 2021-03-31
• Keywords: 情報推薦 / 分散表現 / レビュー分析

Outline of Final Research Achievements

By working on the following two sub-themes, we have developed an information recommendation framework based on item operations that incorporates human feelings toward objects and things. (A) As a feature representation method for items using user reviews, we extended it to various items such as product reviews and music videos to construct a more general model. (B) For information recommendation based on the operation of combining item features, we extended the operation method to various items in FY2020, and also worked on the operation method across items.
As a result, we have presented our work in one journal, five international conferences, and 38 domestic research meetings and received four awards at domestic and international conferences, and have been highly evaluated externally.

Free Research Field

メディア情報学

Academic Significance and Societal Importance of the Research Achievements

情報推薦に関する問題としてフィルターバブルがあげられる．これは，Pariserが主張している，ユーザが接する情報の範囲に対する個人化技術の問題である．推薦アルゴリズムなどの個人化技術によって，知らず知らずのうちにユーザが感心を寄せる特定領域の情報にしかアクセスしなくなる問題である．このような状態になると，感心領域外の新たな情報取得機会が失われるなどユーザの意思決定に対して少なくない影響を与えてしまうことを指摘している．
この問題に対し，本研究成果は(1)推薦結果の多様性に寄与し，(2)推薦根拠が利用者にわかり，(3)ユーザが推薦の方向性を入力可能な形で提供するための知見を得ることに貢献した．