System and method for information retrieval by using keywords associated with a given set of data elements and the frequency of each keyword as determined by the number of data elements attached to each keyword

摘要

The system and method of this invention provides an approach for automatically indicating what property a current set of documents possesses, and an approach for retrieving a set of similar documents by specifying a document rather than a keyword. First, in a set of keywords attached to the set of documents narrowed in response to user retrieval processing, a system according to the present invention counts the number of documents to which each of the keywords is attached to and displays the keywords in order of decreasing frequency. Next, a user specifies the document(s) of interest among the set of documents. In response to this, out of the keywords attached to the documents specified by the user, a keyword is able to retrieve any documents other than the one(s) specified. From the set of keywords attached to the current set of document narrowed, this is determined by omitting the keywords meeting the following conditions: (1) the keyword is not attached to any documents other than the specified one; and (2) the keyword is attached to all the documents in the current set of documents. An automatic retrieval with this keyword will enable narrowing by document rather by the keyword.