Question: 1. The optimization problem solved by lasso regression can be written as: min||X y||22 +||||1 Suppose we increased the size of the (the regularization parameter).
1. The optimization problem solved by lasso regression can be written as: min||X y||22 +||||1 Suppose we increased the size of the (the regularization parameter). Would this likely improve or deteriorate the performance of the model on new data? Why?
2. Suppose we increase the feature representation and training data . Would this likely to increase or decrease the RSS of true error? Why?
Step by Step Solution
There are 3 Steps involved in it
Get step-by-step solutions from verified subject matter experts
