Abstraction of instance space in rule discovery

2018 Fiscal Year Final Research Report

• Project/Area Number: 15K04184
• Research Category: Grant-in-Aid for Scientific Research (C)
• Allocation Type: Multi-year Fund
• Section: 一般
• Research Field: Experimental psychology
• Research Institution: Kindai University
• Principal Investigator: TERAI Hitoshi 近畿大学, 産業理工学部, 教授 (30397442)
• Project Period (FY): 2015-04-01 – 2019-03-31
• Keywords: 規則発見 / 科学的発見

Outline of Final Research Achievements

In this study, we focused on effects of abstraction of instance space in rule discovery. We revealed that (1) abstraction of each instance contributes to find a rule underlying a huge instance space. In particular, (2) focusing on superficial features of each instance was dominant in the early phase of problem solving and decreased toward the middle phase, while focusing on structural features tended to increase from the middle phase to the late phase. Moreover, self-reflection reflecting on their own actions increased at the last phase of problem solving. This implies that self-reflection becomes possible after the abstract understanding of each instance has progressed. Finally, (3) we verified that the abstraction of each instance facilitates searching for the instance space by constructing and simulating a cognitive model for searching for a large amount of an instance space.

Free Research Field

認知科学

Academic Significance and Societal Importance of the Research Achievements

膨大なデータ（事例）の中から法則性を見いだす科学的発見は，科学的活動にとどまらず，学習活動などにも関連する人間の知的活動を支える重要な能力の一つである．事例の抽象化，それに基づく対象に対する理解の促進というサイクルは，現象との相互作用を通して高度な知識構築を実現してきた，化学，生物学，物理学，理工学，そして数学など広く科学一般の活動を支える基礎的なプロセスである．本研究における事例の抽象的理解が規則発見に果たす役割についての実証的知見は，科学的発見に対する基礎的な視点を提供するものである．