人工知能
Online ISSN : 2435-8614
Print ISSN : 2188-2266
人工知能学会誌(1986~2013, Print ISSN:0912-8085)
プロセス,状態,イベントの区別から導かれる事象間の時間関係の生成
東条 敏
著者情報
解説誌・一般情報誌 フリー

1995 年 10 巻 6 号 p. 904-912

詳細
抄録

A legal case consists of a number of temporally entangled affairs. To formalize this temporal structure, we need to write down relations between events and states with much effort, while most of them may be useless. We propose a method to generate these temporal relations, automatically from temporal features of each affair. We first distinguish static affair, processive affair, and events that culminate to the target state. Thereafter we formalize the classification by a unique temporal ontology, that consists of several intervals and a time point, and the notion of perspective given upon the ontology. According to the perspective, that is also some temporal domain in an affair, time intervals are introduced for each affair. We give default rules for aligned sequence of affairs, to relate intervals and affairs. All these temporal relations are formailized in situation thoery ; a situation that is a temporal domain introdued by a preceding affair supports a culmirative point of the succeeding affair. These temporal relations are represented in a formal typed language, and we give a logical model for that language in which we can evaluate truth values of supporting relations of situation theory. A Prolog program is developed to illustrate the feasibility.

著者関連情報
© 1995 人工知能学会
前の記事 次の記事
feedback
Top