PartitionBlock Method of Feature Selection from Information System
-
摘要: 利用差异矩阵和最小差异表,结合与或树技术,提出了一种特征选取的方法IBFS.该方法采用指定标示后分解信息系统为标示块(IB),再根据各个标示块最终求解信息系统的一个较优特征选取.它不仅具有确定性,而且具有快速和有效的特点.Abstract: After utilizing the technique of the combination of And/Or Tree with Different Matrix(DM) and Minimum Different List(MDL), this paper presents a method called IBFS to compute a better feature selection from larger Information System(IS).IBFS mainly partitions larger IS into small ones(called identification block, IB) by the identification feature or features, and then computes the feature selections from IBs, finally forming a feature selection for IS. It has the advantage of high efficiency and determinacy.
-
Key words:
- information systems /
- feature selection /
- heuristic approach
-
[1] Liu H,Dash M. Hybrid search of feature subset. Information Systems and Computer Science Department, National University of Singapore, Technical Report 9/97, 1997 [2] 朱 明,王俊普,蔡庆生.一种最优特征集的选择算法[J].计算机研究与发展,1998,35(9):803~809 Zhu Ming, Wang Junpu,Cai Qingsheng. An algorithm for the optimal feature subsetselection[J].Computer Research & Development,1998,35(9):803~809(in Chinese) [3] 王 珏,王 任,苗夺谦,等.基于Rough Set理论的"数据溶缩"[J].计算机学报,1998,21(5):393~400 Wang Jue,Wang Ren,Miao Duoqian,et al. Data enriching based on rough set theory[J]. Chinese J Computers,1998,21(5):393~400(in Chinese) [4] Skowron A, Rauszer C. The Discernibility matrice and function in information systems,intelligent decision support, handbook of applications and advances of rough sets theory[M].Dprdrecht:Kluwer,1992.331~262 [5] 林尧瑞,马少平.人工智能导论[M].清华大学出版社,1997 Lin Yaorui, Ma Shaoping.Introduction of artificial intelligence[M].Beijing:Tsinghua University Press,1997(in Chinese) -
点击查看大图
计量
- 文章访问数: 2762
- HTML全文浏览量: 152
- PDF下载量: 658
- 被引次数: 0
下载:
