Abstract—Ontology plays a vital role in formulating natural language documents to machine readable form on the semantic web. For ontology construction information should be extracted from web documents in the form of entities and relations between them. Identifying syntactic constituents and their dependencies in a sentence, boost the information extraction from natural language text. In this paper we describe the use of Inductive logic Programming as the learning technique used by a multi agent system to perform relation extraction between two identified entities. The learning capability of agents is exploited to train an agent to learn extraction rules from the syntactic structure of natural language sentences. Typed dependencies of the syntactic constituents of sentences provide the background information for the search space to find ingredients for rule induction. In the multi agent system one agent makes use of Inductive Logic Programming for the rule learning process while another agent is expected to use the learnt rules to identify new relations as well as extract instances of predefined relations. All the relations derived are expressed as predicate expressions of two entities. We evaluate our agent system by applying it on number of wikipedia web pages from the domain of birds.
Index Terms—Ontology, Agent, Parser, Annotations, Tagging, Entities, Relations, Predicate, Atom
M. D. S. Seneviratne is with the Institute of Technology University of Moratuwa, Sri Lanka (e-mail: firstname.lastname@example.org).
D. N. Ranasinghe is with the University of Colombo School of Computing, Colombo, Sri Lanka (e-mail: email@example.com).
Cite: M. D. S. Seneviratne and D. N. Ranasinghe, "Inductive Logic Programming in an Agent System for Ontological Relation Extraction," International Journal of Machine Learning and Computing vol. 1, no. 4, pp. 344-352 , 2011.