Exercise

Generating datasets for classification

Finding an actual dataset meeting all desired combinations of criteria can be complicated and, if collected, have privacy concerns. As a solution, you can use dataset generators to give good approximations of real-world datasets.

In this exercise, you will create a large dataset for a 3-class classification problem. For easy visualization of the generated data as a scatter plot, a custom function has been provided as plot_data_points().

Instructions

100 XP

Import the corresponding function from sklearn.datasets for generating classification datasets.
Generate 5000 samples with 4 features, 1 cluster per class, 3 classes, and a class separation of 2.
Print the shape of the generated data.
See the resulting scatter plot.