Towards Building Contextual Representations of Word Senses Using Statistical Models
Claudia Leacock, Geoffrey G. Towell, Ellen M. Voorhees · 1993
Autom&tlc corpus-based sense resolution, or sense disambiguation, techniques tend to focus either on very local context or on topical context. Both components are needed for word sense resolution. A contextual representation of& word sense consists oftopical context and local context. Our goal is to construct contextual representations by automatically extracting topical and local information from textual corpora. We review an experiment evalugting three statistical classifiers that &utomatically extr&ct topical context. An experiment designed to examine human subject performance with similar input is described. Finally, we investigate & method for automatically extracting local context from a corpus. Preliminary results show improved perfor- mance.