Abstract:The knowledge characteristics weighting plays an extremely important role in the effective and accurate knowledge classification. The present characteristics weighting methods always rely heavily on the experts’ priori knowledge. Rough set weighting method can get rid of priori knowledge so as to reflect the objectivity of the data. For the current rough set weighting methods, they could neither help obtain the weight because of the redundant characteristics, nor could the obtained weight reflect the objective situations. In this paper, a new method based on rough set and knowledge granulation theories was proposed to ascertain the characteristics weight. Experimental results on several UCI data sets demonstrate the effectiveness of this new weighting method, such as its subjective randomness, applicable and explicable.