Lasso for prediction and model selection

Learn about the new features in Stata 16 for using lasso for prediction and model selection. Fit models for continuous, binary, and count outcomes using the lasso or elastic net methods; for continuous outcomes, you can also use the square-root lasso method. Choose from different selection methods, including adaptive lasso and cross-validation. After fitting a lasso model, explore your results with cross-validation function plots and coefficient path plots, and compare the fit across multiple lassos.

This video demonstrates how to fit a linear lasso model, create a cross-validation plot, and assess the goodness of fit. You'll see how to create lists of potential covariates with -vl-, and how to split your data into random samples for use with model selection and testing of predictions.
https://www.stata.com

Видео Lasso for prediction and model selection канала StataCorp LLC