Identifying important features statistically

This method will help you make use of some statistical tests for identifying and selecting relevant features. So, for example, for classification tasks we can use an ANOVA F-statistic to evaluate the relationship between numerical features and the target, which will be a categorical feature because this is an example of a classic task. Or, to evaluate the statistical relationship between a categorical feature and the target, we will use the chi-squared test to evaluate such a relationship. In scikit-learn, we can use the SelectKBest object and we will see how to use these objects in a Jupyter Notebook.