摘要
传统的K-means算法是迭代型算法,它在解决问题的过程中经常得到的是局部的最优解而不是全局的最优解。本文针对这一问题对K-means算法进行改进,改进后的K-means算法对初始点不敏感,且聚类效果好。同时本文将改进后的K-means算法应用到智能用电的数据分析上,通过MATLAB建立了房屋面积、家庭成员、用电量等数据模型,使用改进的K-means算法进行数据挖掘实验,实验结果证明提出的算法稳定可行,且能挖掘出用电数据的潜在的有价值的信息。
he traditional K-means algorithm is iterative algorithm, which often gets the local optimal solution rather than global optimal one in problem solving process. This paper improves the K-means algorithm to address this problem. The improved K-means algorithm is not insensitive to initial point and has good clustering effect. Meanwhile, this paper applys the improved K-means algorithm to the data analysis of Smart power, building the data model of floor area, family members,electricity consumption etc. through MATLAB platform, using the improved K-means algorithm for data mining experiment. Results have shown that the proposed algorithm is stable and feasible and can dig out potentially valuable information of electricity data.
作者
胡涛
王涛
史永帅
鞠明远
HU Tao;WANG Tao;SHI Yong-shuai;JU Ming-yuan
出处
《信息技术与信息化》
2016年第9期39-43,共5页
Information Technology and Informatization
基金
山东省信息产业专项发展资金项目《基于物联网的大型建筑物智能用电及能耗优化关键技术研究》(2012X0107)资助
