Comparison three methods of clustering: k-means, spectral clustering and hierarchical clustering

• Main Author: Kowsari, Kamran
• Format: Journal Article
• Language: English
• Published: 19.12.2013
• Subjects: Computer Science - Learning
• DOI: 10.48550/arxiv.1312.6117

Comparison of three kind of the clustering and find cost function and loss function and calculate them. Error rate of the clustering methods and how to calculate the error percentage always be one on the important factor for evaluating the clustering methods, so this paper introduce one way to calculate the error rate of clustering methods. Clustering algorithms can be divided into several categories including partitioning clustering algorithms, hierarchical algorithms and density based algorithms. Generally speaking we should compare clustering algorithms by Scalability, Ability to work with different attribute, Clusters formed by conventional, Having minimal knowledge of the computer to recognize the input parameters, Classes for dealing with noise and extra deposition that same error rate for clustering a new data, Thus, there is no effect on the input data, different dimensions of high levels, K-means is one of the simplest approach to clustering that clustering is an unsupervised problem.