I've used Ridge, Lasso and Elastic-net regression using different values of α to study how effective they are in cases where the number of observations is very less than the no of predictors. I've created an R code which analyses how the values of MSE change as the values of α and α^2 increase. This repository contains the R code and a PPT which is a step by step guide on achieving the plot. Please go through the notes section of the PPT as it explains step by step how I achieved the plot.
capedcrusader03 / alpha-and-alpha-2-vs-mse-for-elastic-net-regression-in-r Goto Github PK
View Code? Open in Web Editor NEWA study on the effect of MSE on α and α^2 on Elastic-net regression