Analysing High Dimensional Data using Rough Tolerance Relation

It is estimated that there is an exponential growth of data and it is being sprawl in many devices and cloud platforms. Organizing these data in proper pattern is an essential task for data scientists. Dimensionality reduction or removal of inconsistent variables is a major task of organizing high dimensional data. Rough set plays an important role in attribute reduction and it finds hidden patterns from the data without expecting additional parameters. This theory was constructed through indiscernibility relation between objects which is an equivalence relation. In this paper we have opted tolerance relation and propose an intelligent tool through rough hybridization technique with neural network. Rough neural network is an essential branch of granular computing and it was introduced by Pawn Lingras [1]. In this paper we propose tolerance based Rough-Neural network algorithm for attribute reduction and this algorithm is being implemented in SECOM data from UCI repository.