Machine Learning with scikit-learn Quick Start Guide

The optimal number of clusters

While explaining how the k-means algorithm works, we mentioned how the algorithm terminates once it finds the optimal number of clusters. When picking clusters arbitrarily using scikit-learn, this is not always the case. We need to find the optimal number of clusters, in this case.

One way that we can do this is by a measure known as inertia. Inertia measures how close the data points in a cluster are to its centroid. Obviously, a lower inertia signifies that the groups or clusters are tightly packed, which is good.

In order to compute the inertia for the model, we use the following code:

# Inertia of present modelk_means.inertia_

The preceding code produced an inertia value of 4.99 × 10 ^ 17, which is extremely ...