Word Sense Disambiguation With Sublexical Representations
Hinrich Schfitze, Ventura Hall · 1992
This paper introduces a new representational scheme for word sense disambiguation. Drawing on work in information retrieval (Latent Semantic Indexing as proposed by [Deerwester et al. 1990]) an efficient method for learning sublezical rep-resentations is described: Words and contexts are represented as vectors in a multidimensional space which approximates similarity of colloca-tional patterns. Closeness of words in the space is equivalent to occurrence in similar contexts, thus giving a rough approximation of semantic similar-ity. The Bayesian classification system AutoClass was then used to perform an unsupervised classifica-tion of sublexical representations of contexts of the three ambiguous words interest, suit and plant in a training text. In applying this classification to a test text, AutoClass disamhiguated 90 % of all occurrences correctly. Unsupervised classification failed for tank, but a more sophisticated algorithm also achieved a disambiguation rate of 90%.