SYNONYM EXTRACTING METHOD, DOCUMENT RETRIEVING METHOD AND DEVICE TO BE USED FOR THE SAME

摘要

PURPOSE: To enable high-reproducibility document retrieval by expanding a retrieval condition by extracting a synonym or translation concerning the retrieval condition composed of arbitrary character strings. CONSTITUTION: This synonym extracting device is provided with an index preparing means 102 for preparing an index from a document set 101, a document retrieving means 108 for acquiring frequency information by retrieving the index to a retrieval expression composed of arbitrary character strings, a means 110 for preparing a feature vector from the frequency information, a feature vector table preparing means 104 for calculating feature vectors corresponding to all words in the document set and storing them as a table and a similarity calculating means 112 for calculating the similarity of the feature vectors. The feature vector is dynamically prepared from the frequency information, the similarity with the vector on the feature vector table is calculated by the similarity calculating means, words having high similarity between vectors are extracted as synonym and an inquiry is expanded while utilizing the synonym so that document retrieval can be performed with much higher reproducibility.