Feature selection for automatic taxonomy induction

Most existing automatic taxonomy induction systems exploit one or more features to induce a taxonomy; nevertheless there is no systematic study examining which are the best features for the task under various conditions. This paper studies the impact of using different features on taxonomy induction for different types ions and for terms at different abstraction levels. The evaluation shows that different conditions need different technologies or different combination of the technologies. In particular, co-occurrence and lexico-syntactic patterns are good features for is-a, sibling and part-of relations; contextual, cooccurrence, patterns, and syntactic features work well for concrete o-occurrence works well for abstract terms. Categories and Subject Descriptors H.3.1 Content Analysis and Indexing. General Terms Experimentation, Verification. Keywords Ontology Learning, Taxonomy, Semantic Feature.

• Type: Conference
• Year: 2009
• Where: SIGIR
• Authors: Hui Yang, Jamie Callan
• Keywords: Information Retrieval, SIGIR 2009, Syntactic Features, Taxonomy Induction, Taxonomy Induction Systems