Machine Learning for OpenCV

Implementing our own expectation-maximization solution

The expectation-maximization algorithm is simple enough for us to code it up ourselves. To do so, we will define a function find_clusters(X, n_clusters, rseed=5) that takes as input a data matrix (X), the number of clusters we want to discover (n_clusters), and a random seed (optional, rseed). As will become clear in a second, scikit-learn's pairwise_distances_argmin function will come in handy:

In [7]: from sklearn.metrics import pairwise_distances_argmin...     def find_clusters(X, n_clusters, rseed=5):

We can implement expectation-maximization for k-means in five essential steps:

1. Initialization: Randomly choose a number n_clusters of cluster centers. We don't just pick any random number ...