A type of fuzzy information retrieval using a fuzzy thesaurus is appropriate for the similar case retrieval in the case-based reasoning, because we can use index attribute values of the cases as keywords of documents. In this retrieval, the fuzzy thesaurus plays an important role, and hence the method to generate a reasonable fuzzy thesaurus is a problem to be solved. In this paper, a new algorithm to generate a fuzzy thesaurus for the case-based reasoning is presented, which makes use of index attribute values. The index attribute values labeled on cases are electronic data for use of retrieval computation. There is only one index attribute value in each index and the simultaneous occurrence of the index attribute values is not observed consequently. This new algorithm utilizes the index attribute values of multiple indexes and evaluates a correlation between simultaneous occurrences of a pair of index attribute values in different indexes. The correlation composes a matrix, which we call "a cross-index matrix". The fuzzy thesaurus of index attribute values in an index is generated based on the cross-index matrix.
[1]
Janet L. Kolodner,et al.
Case-Based Reasoning
,
1989,
IJCAI 1989.
[2]
Gerard Salton,et al.
Automatic Information Retrieval
,
1980,
Computer.
[3]
Tetsuo Kirimoto,et al.
A document retrieval system based on citations using fuzzy graphs
,
1990
.
[4]
Sadaaki Miyamoto,et al.
Fuzzy Sets in Information Retrieval and Cluster Analysis
,
1990,
Theory and Decision Library.
[5]
Tadeusz Radecki.
Outline of a fuzzy logic approach to information retrieval
,
1981
.