一种高效的不协调决策表约简算法

目前,不协调决策表的分布约简、最大分布约简和分配约简算法复杂度较高,不适合处理大数据集。在分析已有算法基础上,分析了基于相对可区分度的属性重要性度量的性质,解决了正域度量属性重要性的缺陷。针对不协调决策表,给出了多种简化协调决策表的定义,从而大大缩减了约简的实例数。以相对可区分度为启发函数构造了一种高效完备的不协调决策表约简算法。理论分析和实验结果表明,该约简算法解决了现有算法在复杂度和属性重要性度量上的缺陷,适合处理不协调的大数据集。

An efficient attribute reduction algorithm in inconsistent decision tables

Existing algorithms of distribution reduct, maximum distribution reduct and assignment reduct for inconsistent decision tables are inefficient, which are not suitable for large data sets. In this paper, the measure of attribute importance based on the relative discernibility degree is presented firstly, which overcomes the shortcoming of positive domain in measuring the importance of attributes. In addition, in order to simplify decision table, some kinds of simplified consistent decision tables are defined. At last, an efficient attribute reduction algorithm is designed based on relative discernibility degree. Theoretical analysis and experimental results show the effectiveness and practicalbility of this algorithm on the large inconsistent data sets.