Application of rough sets to information retrieval

The aim of the present article is to develop a method of rough retrieval, namely, an application of the rough set theory to information retrieval. After a brief review of fuzzy sets, rough sets, and a fuzzy logical model for information retrieval, rough approximations for retrieved data are defined. The approximations are considered for both crisp and fuzzy cases. A fuzzy set is introduced for the rough boundary, and estimation for the membership for the results of set operations on the boundary is discussed. Rough approximations in cases when hierarchical classes are assumed are considered. Moreover, another approximation by a membership sequence is discussed which refines the foregoing approximations. Illustrative examples are shown.