Methods and apparatus for generating query feedback based on co-occurrence patterns

摘要

An information retrieval system generates query feedback terminology for an input query based on a corpus of documents and the input query. Potential query feedback terms are identified through the corpus of documents for use as query feedback in the information retrieval system. To process a query, which includes at least one query term, co-occurrence signatures for “N” query feedback terms are generated. An N×N matrix of the co-occurrence signatures depict patterns of semantic distance and conceptual proximity among the “N” query feedback terms. The N×N matrix is processed to reduce the number of the entries in the signature to an N×M matrix while preserving the co-occurrence distance characteristics among the signatures. Query feedback terms with co-occurrence signatures, which compare with co-occurrence signatures of the query term in a predetermined manner, are selected as query feedback.