摘要
邻域粗糙集是经典Pawlak粗糙集的扩展,能够有效的处理数值型数据。因为引入了邻域粒化的概念,使用邻域粗糙集模型计算样本邻域度量属性重要度时,需要不断反复的对负域中的样本进行邻域划分操作,算法计算量很大。为此提出了一种基于Relief算法属性重要度的快速属性约简算法,降低计算邻域的算法时间复杂性。通过和现有算法运用多组UCI标准数据集进行比较,实验结果表明,在不降低分类精度的前提下,该算法能更快速地得到属性约简。
Neighborhood rough set is an extension of classical Pawlak rough set, which can deal with nu merical data effectively. However, because the concept of neighborhood granulation is introduced. When the neighborhood rough set model is used to calculate the neighborhood of samples to measure the impor tance of attributes, the neighborhood partition operation of samples in the negative domain should be repeated. Therefore, a fast attribute reduction algorithm based on Relief algorithm is proposed, which reduces the time complexity of calculating the neighborhood. Compared with the existing algorithms using multiple sets of UCI standard data sets, the experimental results show that the algorithm can get attribute reduction more quickly without reducing the classification accuracy.
作者
林芷欣
刘遵仁
纪俊
LIN Zhi-xin;LIU Zun-ren;JI Jun(College of Computer Science and Technology, Qingdao University, Qingdao 266071, China)
出处
《青岛大学学报（自然科学版）》
CAS
2019年第3期8-13,共6页
Journal of Qingdao University(Natural Science Edition)
基金
国家自然科学基金项目(批准号:61503208)资助
关键词
邻域粗糙集
邻域计算
RELIEF算法
属性重要度
属性约简
neighborhood rough set
neighborhood computing
Relief algorithm
attribute significance
attribute reduction
