Hi,
I used a multiple linear regression with R 's lm() function to understand the importance of each regressor thanks to their coefficient / weight. And I want to reproduce it on DSS.
When I create a prediction model with Expert mode and "Choose Algorithms", in "Basic target" I indicate "Regression" for "target prediction type": in "Modeling > Algorithms" only a Ridge regression is proposed , not the equivalent of lm() function in R.
In addition, I don't want a train/test set but full dataset for train (it's for undestanding not to predict).
So is it possible to have 100% dataset for train and use the equivalent of lm() function of R in DSS?
Thanks to you.
Théo
Théo (@titouan3412)
Just a thought. You can add ML Models to Dataiku. One of the easiest ways is to use some of the other models provided by SciKit-Learn in a Custom Model.