融合密度峰值和模糊C-均值聚类算法
|
摘要
针对模糊C—均值(FCM)聚类算法聚类结果依赖于初始中心的选取,易收敛于局部极值等问题,提出了一种密度峰值聚类(DPC)算法和FCM相结合的混合聚类方法(DPC-FCM),利用密度峰值快速搜索算法可以比较准确地刻画聚类初始中心的特点,改善FCM聚类算法存在的不足,从而实现优化聚类。在UCI数据集和人工模拟数据集上的实验结果显示:融合后的新算法和传统的FCM算法相比有着更高的正确率和更快的收敛速度,证明了新算法的可行性。
Aiming at problem that clustering result of fuzzy C-means( FCM) clustering algorithm depends on selection of initial center,easy convergence to local minima,a hybrid clustering method combines density peak fast search algorithm( DPC) with FCM algorithm( DPC-FCM) is proposed. Algorithm uses DPC algorithm can accurately describe characteristics of initial cluster centers,to improve the shortcomings of the FCM clustering algorithm,so as to realize the optimization of clustering. The experimental results of the UCI data sets and artificial data sets show that the fused new algorithm has a higher accuracy and faster convergence speed,than traditional FCM algorithm thus proves the feasibility of the new algorithm.
Aiming at problem that clustering result of fuzzy C-means( FCM) clustering algorithm depends on selection of initial center,easy convergence to local minima,a hybrid clustering method combines density peak fast search algorithm( DPC) with FCM algorithm( DPC-FCM) is proposed. Algorithm uses DPC algorithm can accurately describe characteristics of initial cluster centers,to improve the shortcomings of the FCM clustering algorithm,so as to realize the optimization of clustering. The experimental results of the UCI data sets and artificial data sets show that the fused new algorithm has a higher accuracy and faster convergence speed,than traditional FCM algorithm thus proves the feasibility of the new algorithm.
引文
[1]付争方,朱虹.基于模糊理论的低照度彩色图像增强算法[J].传感器与微系统,2014,33(5):121-124.
[2]张国锁,周创明,雷英杰.改进FCM聚类算法及其在入侵检测中的应用[J].计算机应用,2009,29(5):1336-1338.
[3]Rodriguez Alex,Alessandro Laio.Clustering by fast search and find of density peaks[J].Science,2014,344(6191):1492-1496.
[4]Zelnik-Manor L,Perona P.Self-tuning spectral clustering[J].Advances in Neural Information Processing Systems,2004,17:1601-1608.
[5]李文杰,廖晓纬,束仁义,等.基于Zig Bee和模糊C均值聚类算法在火警中的应用[J].传感器与微系统,2012,31(10):143-152.
[6]杨燕,靳蕃,Kamel M.聚类有效性评价综述[J].计算机应用研究,2008,41(6):1631-1632.
[7]韩晓慧,王联国.一种基于改进混合蛙跳的聚类算法[J].传感器与微系统,2012,31(4):137-139.
[8]高孝伟.权熵法在教学评优中的应用研究[J].中国地质教育,2008,17(4):100-104.
[9]华漫,李燕玲,魏永超.基于自适应相似度距离的改进FCM图像分割[J].电视技术,2016,40(2):33-36.
[2]张国锁,周创明,雷英杰.改进FCM聚类算法及其在入侵检测中的应用[J].计算机应用,2009,29(5):1336-1338.
[3]Rodriguez Alex,Alessandro Laio.Clustering by fast search and find of density peaks[J].Science,2014,344(6191):1492-1496.
[4]Zelnik-Manor L,Perona P.Self-tuning spectral clustering[J].Advances in Neural Information Processing Systems,2004,17:1601-1608.
[5]李文杰,廖晓纬,束仁义,等.基于Zig Bee和模糊C均值聚类算法在火警中的应用[J].传感器与微系统,2012,31(10):143-152.
[6]杨燕,靳蕃,Kamel M.聚类有效性评价综述[J].计算机应用研究,2008,41(6):1631-1632.
[7]韩晓慧,王联国.一种基于改进混合蛙跳的聚类算法[J].传感器与微系统,2012,31(4):137-139.
[8]高孝伟.权熵法在教学评优中的应用研究[J].中国地质教育,2008,17(4):100-104.
[9]华漫,李燕玲,魏永超.基于自适应相似度距离的改进FCM图像分割[J].电视技术,2016,40(2):33-36.