基于K均值、EM和K* 均值算法的团簇行为分析
An Analytical Study on Behavior of Clusters Using K Means, EM and K* Means Algorithm
摘要:
聚类是一种无监督学习方法,构成了智能数据分析过程的基石。它用于探索模式集合之间的相互关系,通过将它们组织成同质集群。聚类已被动态地应用于信息检索(IR)领域的各种任务。集群已经成为当前研究和发展最为活跃的领域之一。聚类试图发现一组相应的组,其中每个组中的组比分配给不同组的组中的组彼此之间的关系更密切。所得到的聚类可以提供用于组织大量文本的结构,以实现高效的浏览和搜索。在聚类问题中,存在着各种各样的聚类算法,这些算法已经得到了深入的研究。在最常用和最有效的算法中,迭代优化聚类算法已经被证明具有合理的聚类性能,例如期望最大化(EM)算法及其变体,以及众所周知的k-means算法。本文分析了分区方法聚类技术- EM,K -均值和K* 均值算法如何在具有以下特征-纯度,熵,CPU时间,聚类分析,均值分析和类间距离的heartspect数据集上工作。最后给出了5个聚类数据集的实验结果,进一步证实了EM算法在聚类质量上远优于k-means算法和k*means算法的结论。
英文摘要:
Clustering is an unsupervised learning method that constitutes a cornerstone of an intelligent data analysis process. It is used for the exploration of inter-relationships among a collection of patterns, by organizing them into homogeneous clusters. Clustering has been dynamically applied to a variety of tasks in the field of Information Retrieval (IR). Clustering has become one of the most active area of research and the development. Clustering attempts to discover the set of consequential groups where those within each group are more closely related to one another than the others assigned to different groups. The resultant clusters can provide a structure for organizing large bodies of text for efficient browsing and searching. There exists a wide variety of clustering algorithms that has been intensively studied in the clustering problem. Among the algorithms that remain the most common and effectual, the iterative optimization clustering algorithms have been demonstrated reasonable performance for clustering, e.g. the Expectation Maximization (EM) algorithm and its variants, and the well known k-means algorithm. This paper presents an analysis on how partition method clustering techniques - EM, K -means and K* Means algorithm work on heartspect dataset with below mentioned features - Purity, Entropy, CPU time, Cluster wise analysis, Mean value analysis and inter cluster distance. Thus the paper finally provides the experimental results of datasets for five clusters to strengthen the results that the quality of the behavior in clusters in EM algorithm is far better than k-means algorithm and k*means algorithm.
机构、总结、中文摘要均由AI分析而来,可能存在错误,请以论文正文为准。