摘要
约简是粗集理论的重要概念,由定义计算约简是一个典型的NP问题且由于约简的不唯一,在面对大数据集或高维数据集问题时获得的属性集往往并非是最小的属性约简集。文中针对Rough sets理论的属性约简进行了研究。研究了通过可辨识矩阵求得属性约简集,利用Rough sets与灰色理论相结合,提出一种属性约简的启发式算法,拟合结果表明本约简算法合有效。
Reduction is an important concept in rough set theory, while computing reduction according to the definitions directly is a typical NP problem. The attribute set get from the problem with large and high- dimension database is not usually the minimum attribute set. Discusses the approaches for attribute reduction based on rough set theory. Following, studies the approaches to achieve attribute reduction set by applying recognized matrix. Researched how to get attribute reductions through discernibility matrix, combined rough sets and gray theory,and put forward a new heuristics algorithm for attribute reduction. The effectiveness of the result obtained is demonstrated by an example.
出处
《计算机技术与发展》
2008年第1期154-156,共3页
Computer Technology and Development
基金
江苏省教育资助项目(2005DX006J)
