An Accuracy Measure for Rough Sets Based On Knowledge of Particles

In rough sets theory, the accuracy quantification is caused by its boundary region. However, Z. Pawklak proposed the traditional accuracy measure which does not take into consideration the knowledge of particles of the partition induced by an equivalence relation. Paper five proposed an improved accuracy measure, which does not consider the large of universe, when the partition is finest, the accuracy measure becomes biggest suddenly. This paper not only take into consideration the large of knowledge of particles but also the large of universe and proposes a new kind of accuracy measure. Meanwhile we give some properties of it. It is more reasonable and effective to measure, which are illustrated by examples.