Variable Precision Rough Set Model and Application Based on Grey Similarity Incidence Relationship

摘要

There exists a large amount of noise data and continuous attribute values in the real decision making information system. Up to now a good amount of the valuable information has resulted from data discretization based on traditional methods. In order to effectively deal with the problem, grey incidence analysis methodology is exploited to construct a novel variable precision rough set model. In this paper, to begin with, the concept of grey similarity incidence is employed to define the grey similarity incidence relationship and grey similarity incidence class with threshold value lambda What's more, grey similarity incidence relationship is utilized to replace the equivalent relationship and then the variable precision rough set model is developed. After investigating the properties of our new model, a case study confirms the validity and rationality of the method by comparing the proposed model with other data discretization methods and variable precision rough set models.