My dataset contains some stocks and some bitcoin titles. I would like to replace NA data (normal stocks are not traded on sat. and sun.) using a regression on the stock itself. How can i do this?
1
-
$\begingroup$ What you try exactly to do and what way/software you try to accomplish the imputation ? How does your data look ? $\endgroup$– n1tkOct 6, 2018 at 1:42
Add a comment
|
1 Answer
$\begingroup$
$\endgroup$
Set the column with missing values as your target.
Remove rows where your target is null.
Train a regression model that predicts your target using the rest of your features.
Replace missing values with this regression model's predictions.
