Learning temporal interval relations using inductive logic programming

Abstract

The notion of time permeates every single aspect of the world around us and, as such, it should be taken into account when developing automatic systems that implement many of its processes. In the literature several proposals for representing the notion of time can be found. One of the most popular is the Allen's temporal interval, based on a set of 13 relations that may hold between two time intervals. The main goal of this work is to explore the automatic learning of several of temporal relations from data, using an inductive logic programming (ILP) system. The paper describes a set of automatic learning experiments whose main aims are (i) determining the impact of the negative training patterns on the induced relation (ii) evidencing the necessary background knowledge for inducing the exact expression of the target concept and (iii) investigate the viability of ILP as a learning mechanism in real-time systems. © Springer-Verlag Berlin Heidelberg 2011.