Combining outputs of ridge regression models?


I am facing an issue where I have 7 sets of different variables/columns/predictors.

I am trying to predict same target variable and I want to observe the importance/effect of all the sets according to their importance in an ordered manner. (I am trying to use ridge regression models for each of the 7 individual set as I want to keep all the variables and I want to combine the output of these 7 models, each set has more than 20 variables)


Posted 2018-08-21T18:39:57.203

Reputation: 13

Why don't you simply rank your set of features based on how good your model is when trained by each set? – Valentin Calomme – 2018-08-22T06:54:24.750

I am looking at a business case where the order of importance of the set of predictors is pre-specified. @ValentinCalomme – PruthvijThakar – 2018-08-22T15:38:31.407



If you mean predictors by "variables", You may combine your 7 models by using "rbind or cbind" and save it into a data frame or data table. But, remember to first save your each of the 7 models into a datatable or dataframe.

About ranking, I agree with @Valentin you may want to make the predictions first, then rank it . However , if you want to take all the predictors into account there is no point to apply the ranking of all the sets. You will see their importance once you are done predicting.

Additional note: Glmnet is way more sensitive library/method when performing predictions, make sure all the columns while creating a model and doing a prediction are in the same order and have the same names. This caused me much trouble while predicting. Hopefully, this might help you.


Posted 2018-08-21T18:39:57.203

Reputation: 134