K-Means clustering is a popular cluster analysis method. It is simple and its implementation does not require to keep in memory all the dataset, thus making it possible to process very large databases.
This course material describes the algorithm. We focus on the different extensions such as the processing of qualitative or mixed variables, fuzzy c-means, and clustering of variables (clustering around latent variables). We note that the k-means method is relatively adaptable and can be applied to a wide range of problems.
Keywords: cluster analysis, clustering, unsupervised learning, partition method, relocation
Slides: K-Means clustering
References :
Wikipedia, "k-means clustering".
Wikipedia, "Fuzzy clustering".
