Discussion and Suggestions for Further Studies

Abstract

We have seen in the foregoing chapters how fuzzy set theory provides various mathematical models for dealing with problems in information retrieval. A simple method of fuzzy information retrieval was introduced first. Then, in Chapter 5, fuzzy retrieval through fuzzy thesauri was considered as a natural generalization of current methods of retrieval through crisp thesauri, and fuzzy associations were shown capable of incorporating methods in bibliographic analysis. Chapter 6 dealt with hierarchical clustering and discussed equivalence among the single linkage method, maximal spanning trees, the transitive closure of a symmetric fuzzy relation, and connected components of a fuzzy graph. A diagrammatic representation of information retrieval process was discussed in Chapter 7 and equivalence between feedback in a diagram and clustering of terms or documents was proved. In Chapter 8 the concept of fuzzy profiles was proposed and a fuzzy set model was applied to evaluation measures of retrieval effectiveness. Also, graphical representation among information units in research literature was considered using fuzzy sets. Running through all these topics is the theme that diverse methods currently used in information retrieval and bibliographic analysis can be put into the unifying framework of fuzzy set theory, and that relationships between these methods can be clarified within this framework.