论文部分内容阅读
为讨论差转计算在多因素决策问题中对因素系统的约简性能,通过理论分析与实证检验,在算法原理与机制的基础上深入讨论了算法的因素约简能力,并与基于差别矩阵的粗糙集因素约简算法进行比较.研究结果表明:差转计算在进行决策的过程中,确保了对结果有重要影响的因素进入经验推理系统,影响不大的因素在决策的过程中自动地被舍弃,从而实现了因素的约简.在6个UCI数据集上,通过对因素约简结果与时间复杂度两个方面的讨论,得出差转计算的因素约简能力同差别矩阵算法相当,在时间复杂度方面远优于差别矩阵算法.
In order to discuss the reduction performance of the factor system in the multi-factor decision making problem, this paper discusses the factor reduction capability of the algorithm based on the principle and mechanism of the algorithm through theoretical analysis and empirical test, and compares it with that based on discernibility matrix The results show that in the process of decision-making, transfer-difference calculation ensures that factors that have an important influence on the result enter into the system of experience-based reasoning, and the factors that have little effect on the decision-making process are automatically Discarding, so as to realize the reduction of the factors.On the six UCI datasets, through the discussion of the factor reduction results and the time complexity, we conclude that the factor reduction ability of the difference calculation is equivalent to the difference matrix algorithm, The time complexity is far superior to the difference matrix algorithm.