scikit-learn : Machine Learning Simplified

Clustering handwritten digits with k-means

K-means is the most popular clustering algorithm, because it is very simple and easy to implement and it has shown good performance on different tasks. It belongs to the class of partition algorithms that simultaneously partition data points into distinct groups called clusters. An alternative group of methods, which we will not cover in this book, are hierarchical clustering algorithms. These find an initial set of clusters and divide or merge them to form new ones.

The main idea behind k-means is to find a partition of data points such that the squared distance between the cluster mean and each point in the cluster is minimized. Note that this method assumes that you know a priori the number of clusters ...