Premium
Representation and compilation of knowledge in point‐based temporal system
Author(s) -
Loganantharaj Rasiah
Publication year - 1991
Publication title -
international journal of intelligent systems
Language(s) - English
Resource type - Journals
SCImago Journal Rank - 1.291
H-Index - 87
eISSN - 1098-111X
pISSN - 0884-8173
DOI - 10.1002/int.4550060506
Subject(s) - computer science , temporal database , knowledge base , redundancy (engineering) , temporal logic , knowledge representation and reasoning , graph , theoretical computer science , artificial intelligence , data mining , operating system
Abstract We propose a temporal model based on time points which allows one to incorporate uncertain temporal relations, absolute sense of time, and partial temporal knowledge. It also maintains the status history and is persistent with respect to its present situation. In this article, we describe how to maintain, update, and retrieve data from this point‐based temporal knowledge base. Temporal knowledge is acquired through reified sentences and temporal ordering relations. the temporal knowledge is maintained in a directed graph called Minimal Temporal Network , which is represented in a matrix called Ordering Matrix . This graph is minimal because any redundancy, if there exists any, is eliminated at the update time. to maintain complete world knowledge of past and present, this temporal system works in conjunction with a situational database. A simple deduction system can reason with the situational database and the ordering matrix. Temporal ordering matrix can be constructed in O( kN 2 ) time and any new constraint can be propagated in O( N 2 ) time where N is the number of transitions and k is the average number of out branching factor of the temporal network. With our approach, most of the temporal queries can be answered in a constant time.