Question: 2. (Linear regression) Recall in linear regression, we have data pairs (X1,y1),(Xn,Yn) and we solve the following objective: ^+argRvminXY22. Here X,Y are the compact form
2. (Linear regression) Recall in linear regression, we have data pairs (X1,y1),(Xn,Yn) and we solve the following objective: ^+argRvminXY22. Here X,Y are the compact form to represent the data: XRnp with its i-th row vector to be Xi. and Y is a vector where its i-th coordinate is Yi=yi. (a) (6') Solve ^. (You should represent ^ in the form of X and Y.) (b) (10) Recall the standard model (with fixed design): Y=X+ E, where N(0,2I), is the ground-truth model, and X is considered as fixed. Calculate the following excess risk of the learned model ^: ER(^)=n1EX^X22. Hint: You may want to use the fact that Tr(AB)=Tr(BA). You msy assume XX is full rank
Step by Step Solution
There are 3 Steps involved in it
Get step-by-step solutions from verified subject matter experts
