Clustering can be considered the most important unsupervised learning problem as clustering deals with unlabeled data. So, as every other problem of this kind, it deals with finding a structure in a collection of unlabeled data. A loose definition of clustering could be the process of organizing objects into groups whose members are similar in some way. A cluster is therefore a collection of objects which are similar between them and are dissimilar to the objects belonging to other clusters. If the similarity criterion is distance: the two or more objects belong to the same cluster if they are close according to
a given distance (geographical distance in case of wireless sensors networks).
There are several ways to cluster a data set.
Moreover, all methods are capable of performing their task for multivariable
data set. While handling such a problem, every method uses its algorithm.
Following are different algorithms which are also the names of the method.
o Prototype based (k-means, k-medoid)
o Linkage Methods (single link, complete link)
o Graph Theoretic (MST, spectral clustering)
• Model based
o Spatial Clustering
o Mixture Model (Gaussian Mixture)
• Density Based
o Mode seeking (means shift)
o Kernel Based (DENCLUE)
o Grid Based (wave-Cluster, STING)
All the listed algorithms are different approaches for an unsupervisedprocedure. Moreover, every algorithm has a criterion function which they have to optimize.