Abstract: This paper discusses a deviation problem of measuring case similarity, which happens in the process of features reduction with classical rough set theory. In order to improve reduction accuracy and keep good quality of reduced case base, this paper proposes an algorithm based on similarity rough set theory to reduce features and compute features weight.

摘要： 讨论了经典粗糙集理论在特征项约简过程中,造成的范例相似度测量误差问题,并提出用相似粗糙集约简特征项和获取特征项权重的过程与算法,力求提高约简的精确性和保持约简后范例库的良好性能。

关键词: 相似粗糙集, 范例推理, 特征项约简, 差别矩阵, 特征项权值

CLC Number: 

• TP301