Investigating the Effectiveness of Thesaurus Generated Using Tolerance Rough Set Model

We considered the tolerance matrix generated using tolerance rough set model as a kind of an associative thesaurus. The effectiveness of the thesaurus was measured using performance measures commonly used in information retrieval, recall and precision, where they were used for the terms rather than documents. A corpus consists of keywords defined as highly related with particular topic by human experts become the ground truth of this study. Analysis was conducted based on comparison values of all available sets created. Above all findings, this paper was thought as the fundamental basis that generating an automatic thesaurus using rough sets theory is a promising way. We also mentioned some directions for future study.