Python Natural Language Processing

Feature selection techniques

Make everything as simple as possible but not simpler

This quote by Albert Einstein, is very true when we are talking about feature selection techniques. We have seen that to get rid of the curse of dimensionality, we need feature selection techniques. We will look at the following feature selection techniques:

• Filter method
• Wrapper method
• Embedded method

So, let's begin with each method.

Filter method

Feature selection is altogether a separate activity and independent of the ML algorithm. For a numerical dataset, this method is generally used when we are preprocessing our data, and for the NLP domain, it should be performed once we convert text data to a numerical format or vector format. Let's first see the ...