Machine Learning Techniques for Automatic Ontology Construction from Domain Texts

Dr. Jianhua Chen

Computer Science Dept., Louisiana State University

Friday, October 12, 2007

3:00 p.m. - 4:00 p.m.

Woodward Hall, Room 106

Complete Description:

Automatic construction of ontologies from texts has become an important area of research in Computer Science and Information Technology, driven by the explosive growth of the Internet and World Wide Web. Ontologies capture critical semantic information of an application domain and represent such information in formal systems, allowing for efficient and automatic inferencing with such domain knowledge. Ontologies have been widely used in applications such as information retrieval, software engineering, knowledge management, intelligent query-answering. The ability to automatically build ontologies from domain texts would greatly facilitate such applications. Current approaches to automatic ontology construction still suffer from a number of limitations in terms of efficiency and quality of ontologies extracted. In this talk, we present techniques that combine Machine Learning, text mining and information retrieval for automatic ontology construction from domain texts. We address three important components of ontology construction: Concept extraction, taxonomy relation extraction and non-taxonomy relation extraction. In concept extraction, we propose to use a combination of information retrieval technique with Wordnet, a general-purpose lexical system to obtain high precision and recall of the target concepts. In taxonomical relation learning, we study the problem of Semantic Class Labeling and propose a machine learning solution to the problem. Compositional structures of phrases are also exploited to learn taxonomical relations. We investigate the use of subject-verb-object triples with a statistical metric for non-taxonomical relation extraction. Results of empirical studies with texts from two domains are also presented. Bio: Dr. Jianhua Chen received her Ph.D. in Computer Science in 1988 from Jilin University, Chang Chun, China. In fall 1988, Dr. Chen joined the Computer Science Department of Louisiana State University, Baton Rouge, USA, where she is currently an Associate Professor. Dr. Chen was an ASEE/NAVY summer faculty research fellow in summer 2002 at US Naval Research Laboratory in NASA Stennis Space Center, Mississippi. Dr. Chen's research interests include machine learning and data mining, Web mining, fuzzy Logic and Fuzzy Clustering, Knowledge Representation and Reasoning. Dr. Chen has published over 100 refereed journal and conference papers, and her research has been supported by National Science Foundation, Naval Research Laboratory, and Louisiana Board of Regents.