Rough sets in hybrid soft computing systems

Abstract

Son computing is considered as a good candidate to deal with imprecise and uncertain problems in data mining. In the last decades research on hybrid son computing systems concentrates on the combination of fuzzy logic, neural networks and genetic algorithms. In this paper a survey of hybrid son computing systems based on rough sets is provided in the field of data mining. These hybrid systems are summarized according to three different functions of rough sets: preprocessing data, measuring uncertainty and mining knowledge. General observations about rough sets based hybrid systems are presented. Some challenges of existing hybrid systems and directions for future research are also indicated. © Springer-Verlag Berlin Heidelberg 2007.