摘要
针对传统k-均值聚类算法中每个属性聚类作用相同而导致的聚类效果不佳,以及不适宜在传感器网络中使用等问题,在传感器网络中采用粗糙k-均值算法对数据进行分布式聚类,可减少网络负载和传感器节点能量的消耗.实验结果证明:该算法在聚类速度、聚类正确率、网络传输通信量等方面均优于传统k-均值算法.
In order to solve the problem that the sameness of every attribute clustering effect leads to traditional k- means clustering algorithm being not good and suitable used in the distributed environment, a k-means distributed clustering algorithm based on rough set is proposed. This algorithm adopts rough set attribute reduction algorithm to delete the redundant attributes of clustering sample, which can reduce the data dimension, then determine weights based on the significance of attribute. On the basis of improving the traditional k-means algorithm and getting rough k-means algorithm, and then each node making full use of their own resources makes distributed clustering based on rough k-means. The experimental results demonstrate the superiority of the proposed algorithm.
出处
《广西工学院学报》
CAS
2013年第1期89-93,共5页
Journal of Guangxi University of Technology
基金
广西自然科学基金青年项目(2012jjBAG0074)
广西混杂计算与集成电路设计分析重点实验室开放基金课题(2012HCIC05)资助
关键词
K-均值
粗糙集
无线传感器网络
分布式聚类
k-means algorithm
rough set theory
wireless sensor network
distributed clustering
