Machine Learning for OpenCV 4 - Second Edition

Implementing our first k-means example

First, let's generate a 2D dataset containing four distinct blobs. To emphasize that this is an unsupervised approach, we will leave the labels out of the visualization:

1. We will continue using matplotlib for all of our visualization purposes:

In [1]: import matplotlib.pyplot as plt...     %matplotlib inline...     plt.style.use('ggplot')

2. Following the same recipe from earlier chapters, we will create a total of 300 blobs (n_samples=300) belonging to four distinct clusters (centers=4):

In [2]: from sklearn.datasets.samples_generator import make_blobs...     X, y_true = make_blobs(n_samples=300, centers=4,...                            cluster_std=1.0, random_state=10)...     plt.scatter(X[:, 0], X[:, 1], s=100);

This will generate the following ...