A novel soft set approach in selecting clustering attribute

摘要

Clustering is one of the most useful tasks in data mining process for discovering groups and identifying interesting distributions and patterns in the underlying data. One of the techniques of data clustering was performed by introducing a clustering attribute. Soft set theory, initiated by Molodtsov in 1999, is a new general mathematical tool for dealing with uncertainties. In this paper, we define a soft set model on the equivalence classes of an information system, which can be easily applied in obtaining approximate sets of rough sets. Furthermore, we use it to select a clustering attribute for categorical datasets and a heuristic algorithm is presented. Experiment results on fifteen UCI benchmark datasets showed that the proposed approach provides a faster decision in selecting a clustering attribute as compared with maximum dependency attributes (MDAs) approach up to 14.84%. Furthermore, MDA and NSS have a good scalability i.e. the executing time of both algorithms tends to increase linearly as the number of instances and attributes are increased, respectively.