In this machine learning tutorial we're going to look at another method of automatic feature selection. However, we aren't finished yet...
We need to see which features have been selected or which are the ones that remained after applying the automatic feature selection method.
To do this, we'll have to apply a Boolean mask over the features. What does this mean actually?
Well, Boolean refers to True or False, and in this case, True is represented by 1, while false is represented by 0. So, applying this mask over the features, after the automatic feature selection method will return an array of all the features
and each of these features is going to be replaced by a 0 or a 1.
If the feature is replaced by 0, it simply means that our feature selection method did not select it, so it took it out of the 'equation' so to speak. If the feature was replaced by 1, then the feature remained after the selection.
For this purpose, we're going to use numpy and matplotlib. Once we do that and we get an idea of the features we have after the selection, we'll simply train a LogisticRegression. But we'll train it on both training sets: the one will all the features, and the one with only the features that remained after the automatic feature selection has been applied.
Then, we'll be able to compare whether or not the algorithm performs better with automatic feature selection. Please watch the video for the full walk-through:
To stay in touch with me, follow 
Cristi Vlad Self-Experimenter and Author