K-means算法复杂度的能量分析方法

Energy Analysis Method Of Computational Complexity For K-means

传统的算法复杂度分析方法重点对算法的内在流程进行分析,而且对于数据挖掘时具有很大不确定性,执行时间不能准确确定。用物理学原理从能量的角度对数据挖掘中的经典聚类算法K-means进行分析,并从算法的本质上推导出其复杂度下限,推导出该算法能优化到的极限值。

The traditional method of complexity analysis focus on the inner flow of algorithm.For the intelligence algorithm in data mining,there is some uncertainty when excuting.And the time of excution could not be exanctly known.According to the principles of physics,this paper analysis the K-means,an classical cluster algorithm in data mining.We deduce the lower bound of its complexity from the essence of algorithm,and point out the best optimized value.