A parallel method for computing rough set approximations

Massive data mining and knowledge discovery present a tremendous challenge with the data volume growing at an unprecedented rate. Rough set theory has been successfully applied in data mining. The lower and upper approximations are basic concepts in rough set theory. The effective computation of approximations is vital for improving the performance of data mining or other related tasks. The recently introduced MapReduce technique has gained a lot of attention from the scientific community for its applicability in massive data analysis. This paper proposes a parallel method for computing rough set approximations. Consequently, algorithms corresponding to the parallel method based on the MapReduce technique are put forward to deal with the massive data. An extensive experimental evaluation on different large data sets shows that the proposed parallel method is effective for data mining. (C) 2012 Elsevier Inc. All rights reserved.