摘要
结合胡可云算法中对属性重要性的描述,以条件属性在可辩识矩阵中出现的频率作为启发信息,提出一种基于变精度粗糙集模型的属性约简算法。实验证明,该方法能够有效地对决策表进行属性约简,并具有一定的抗噪声能力。
Based on the description of the attribute importance in Hu Ke-yun＇s algorithm, proposes an attribute reduction algorithm in variable precision rough set model. The algorithm is on the basis of the attribute frequency in the discernibility matrix. An instance indicates that the decision table can be reduced effectively and that the algorithm has certain fault-tolerant ability.
出处
《现代计算机》
2007年第12期17-19,共3页
Modern Computer
关键词
变精度粗糙集
属性约简
属性重要性
可辩识矩阵
属性频率
Variableprecision Rough Set
Attribute Reduction
Attribute Importance
Discernibility Matrix
Attribute Frequency
