Machine Learning for OpenCV

Generating the dataset

As we can now recite in our sleep, a binary classification problem has exactly two distinct target labels (n_classes=2). For the sake of simplicity, let's limit ourselves to only two feature values (n_features=2; for example, an x and a y value). Let's say we want to create 100 data samples:

In [1]: from sklearn import datasets...     X, y = datasets.make_classification(n_samples=100, n_features=2,...                                         n_redundant=0, n_classes=2,...                                         random_state=7816)

We expect X to have 100 rows (data samples) and two columns (features), whereas the vector y should have a single column that contains all the target labels:

In [2]: X.shape, y.shapeOut[2]: ((100, 2), (100,))