I have a numerical feature called $x$ and a categorical feature called $y$. $y$ is an ordinal feature (A,B,C,D,E,F). I am using label encoding for my y feature and when I am seeing the correlation using matrix heatmap it shows that my $x$ feature has 0.95 correlation with $y$. I am using logistic regression algorithm. Should I drop one of them?
1 Answer
Yes, you should drop one of them. It is not a good idea to have highly correlated features in a logistic regression model. You should be able to see the model's accuracy improve after the drop.
-
$\begingroup$ Your answer could be improved with additional supporting information. Please edit to add further details, such as citations or documentation, so that others can confirm that your answer is correct. You can find more information on how to write good answers in the help center. $\endgroup$– Community BotDec 13, 2022 at 15:49