摘要
知识粒度是人工智能领域的一个研究热点。针对粗糙集的不完备决策表,提出一种不需要求出差别矩阵而直接计算属性频率的方法,以此为启发信息改进基数排序算法。在知识粒度领域给出一个快速计算属性约简的启发式算法,其时间复杂度为O(|C|2|U'|)。最后通过实例说明该算法的有效性。
In artificial intelligence field, knowledge granularity is a focus of research. For incomplete decision tables of the rough set, we give a formula, which calculates the attribute frequency directly without obtaining the discernibility matrix, This is then used as the heuristic information to improve the cardinal number sorting algorithm, and in the field of knowledge granularity we give a heuristic algorithm for quick- ly calculating the attribute reduction, which has the time complexity of 0（ ｜ C ｜2｜U＇ ｜ ）. In end of the paper, the validity of the algorithm is il- lustrated by an example.
出处
《计算机应用与软件》
CSCD
北大核心
2013年第7期43-45,50,共4页
Computer Applications and Software
基金
国家自然科学基金项目(60963008)
广西自然科学基金项目(2011GXNSFA018163)
关键词
粗糙集
不完备决策表
知识粒度
差别矩阵
属性约简
Rough set Incomplete decision tables Knowledge granularity Discernibility matrix Attribute reduction
