Computing Term Translation Probabilities with Generalized Latent Semantic Analysis

Term translation probabilities proved an effective method of semantic smoothing in the language modelling approach to information retrieval. We use Generalized Latent Semantic Analysis to compute semantically motivated term and document vectors. Normalized cosine similarity between term vectors is used as term translation probability. Our experiments demonstrate that GLSA-based term translation probabilities capture semantic relations between terms and improve performance on document classification.

• Type: Conference
• Year: 2006
• Where: EACL
• Authors: Irina Matveeva, Gina-Anne Levow
• Keywords: EACL 2006, GLSA-based Term Translation, Natural Language Processing, Term Translation Probabilities, Term Translation Probability