A new algorithm on attribute reduction with fuzzy rough set method

Today, with the development of computer technologies, we encounter with high dimensional databases. The process of removing the unnecessary data form databases, finding the patterns inside the data and using the obtained knowledge is called knowledge discovery in databases. Attribute reduction is defined as reducing of attribute space dimension according to certain criteria. In this context, a lot of studies have been carried out in literature. In these studies, the most interesting one is the attribute reduction with hybridization of fuzzy and rough sets.In this thesis, a new approach is proposed for attribute reduction based on discernibility matrix using the type-2 fuzzy sets. Proposed algorithm is compared with the other attribute reduction algorithms in the literature using the data sets taken from UCI machine learning repository and in order to test the effectiveness of the proposed algorithm, we made use of classification algorithms.