| 摘要 |
PROBLEM TO BE SOLVED: To provide a text retrieving device capable of highly accurately retrieving text information by more suitably calculating the importance of each word. SOLUTION: Retrieving processing is applied to text information stored in a text information storage part 1, a word extraction processing part 2 extracts a word appearing in a seed text information set specified by a retriever, a document frequency counting processing part 3 counts up the number of text information elements in the seed text information set in which the extracted word appears and the number of text information elements in a text set to be retrieved as document frequency in each text information set, and a document frequency weighting processing part 4 weights at least either one of the document frequency of the seed text information set and that of the text information set to be retrieved. A word importance calculation processing part 5 calculates the importance of the word on the basis of the document frequency or its weighted value and a text association degree calculation processing part 6 calculates the degree of association of the seed text information with each text information included in the text set to be retrieved based on the importance.
|
