An Analysis of Rough Set-Based Application Tools in the Decision-Making Process

The rough set theory is one of various methods that are frequently used by researchers in the analysis of complex data to solve different types of problems. Thus, a number of application software and methods have been proposed and published to make use of the benefits of the rough set theory. However, it is quite difficult for a non-rough set expert without any basic knowledge and information to understand and identify the best method or application software. Therefore, this paper proposes to assist the decision maker in selecting the best rough set-based application tool by analysing the capability of several rough set-based application tools in making good decisions. Four rough set-based application tools were selected to deal with the classification problem in the experimental tasks. The tools were ROSE2, 4eMKa2, JAMM and jMAF. The experimental results showed that JAMM, ROSE2 and jMAF returned quite significant results in the classification process. However, the 4eMKA2 performed well in comparison to the other selected software. The validation results of the random forest (RF), support vector machine (SVM) and neural network (NN) also indirectly proved that the dominance-based rough set approach (DRSA) is one of the best approaches to be used in decision-making processes, especially in the classification process.