METHOD AND SYSTEM FOR IDENTIFYING SET OF QUERY-RELEVANT KEYWORDS COMPRISED IN ONE OR MORE DOCUMENTS OBTAINED FROM QUERY, WHICH NEED NOT COMPRISE THE KEYWORDS

摘要

PPROBLEM TO BE SOLVED: To provide a system and method for identifying query-related keywords in documents found in a search using latent semantic analysis. PSOLUTION: A term-weight matrix M comprising one or more document term-weight vectors d is created. Each document term-weight vector d comprises information on frequency in one or more documents obtained from terms in a query. An expanded query term-weight vector qSBexpanded/SBis created from a query term-weight vector q and the term-weight matrix M. With the expanded query term-weight vector qSBexpanded/SBand the document term-weight vectors d, a set of keywords identified as terms related to the query and also comprised in at least one of the documents is located. The query need not comprise the keywords. PCOPYRIGHT: (C)2006,JPO&NCIPI