Learning for Text Categorization and Information Extraction with ILP

Abstract

Text Categorization (TC) and Information Extraction (IE) are two important goals of Natural Language Processing. While handcrafting rules for both tasks has a long tradition, learning approaches used to gain much interest in the past. Since in both tasks text as a sequence of words is of crucial importance, propositional learners have strong limitations. Although viewing learning for TC and IE as ILPproblems is obvious, most approaches rather use proprietary formalisms. In the present paper we try to provide a solid basis for the application of ILPmetho ds to these learning problems. We introduce three basic types (namely a type for text, one for words and one for positions in texts) and three simple predicate definitions over these types which enable us to write text categorization and information extraction rules as logic programs. Based on the proposed representation, we present an approach to the problem of learning rules for TC and IE in terms of ILP.We conclude the paper by comparing our approach of representing texts and rules as logic programs to others.