Clustering algorithms overview

clusteringalgorithms.md

Name	Definition	Example	Characteristics	Organization
Connectivity Models	Data points closer in data space are more similar than those far away	hierachical cluster	easy to interpret but do not scale well	Hierachical
Centroid models	iterative where similarity is intepreted as proximity of data point to centroid	K-means	provide final number of cluster	Non-Hierachical
Distribution Models	Based on probability of data points in a cluster belonging to the same distribution	EM-Algorithm (Expectation-Maximization)	frequent problems of overfitting	Non-Hierachical
Density Models	Isolate different density regions as basis for clustering	Density-Based Clustering of Application with Noise (DBSCAN)	Not good on high dimensional data or clusters with varying densities	Non-Hierachical