欢迎您！东篱公司
退出
-
申报数据库
-
立项数据库
-
成果数据库
-
项目获奖数据库
中文摘要：
针对目前基于决策表的属性约简矩阵算法效率不理想的情况，首先引入了简化决策表的概念，有效剔除了决策表中大量重复冗余的对象，并给出了正区域模型下简化矩阵的构造，从而有效地缩小了求解属性约简的搜索空间；然后从理论上详细分析了基于简化矩阵的属性约简与基于正区域的属性约简是一致的，并给出了一种有效的属性重要性度量方法．为此，在简化矩阵的基础上设出了一种快速的属性约简矩阵算法，最后通过实例分析和实验对比说明了算法的有效性和可行性．
英文摘要：
At present, the efficiency of matrix algorithm for attribute reduction is not ideal. Aiming at this situation, in this paper, firstly the concept of the simplified decision table is introduced ; a large number of repeated objects in decision table are deleted effec- tively. At the same time, the construction of the simplified matrix based on the model of positive region is provided, which can re- duce the searching space of attribute reduction. Furthermore, it is analyzed theoretically that the attribute reduction definition based on simplified matrix is equal to the attribute reduction definition based on positive region. And an efficient measurement method for at- tribute importance is presented. On this condition, a matrix algorithm for quick attribute reduction is designed. Finally, an example and experimental comparisons are employed to illustrate the efficiency and feasibility of the algorithm.
同期刊论文项目
同项目期刊论文
期刊信息
位置：