Linear Regression (25 pts ): Consider fitting a linear regression model for the following data : X -1 0 2 y -1 Using the minimum square loss , 2 N 1 Elf(x)- x/2 , as the loss function, (a) (4 pts ) Fit f (x )= (intercept only mode ), find the best / using the minimum square loss . (4 pts ) (b) (6 pts ) Fit f (x)= ( linear regression within intercept ), fin the best , using the minimum square loss . (c) (7 pts ) Is the model f (x=B + 3 x with the best B and B, from above the best model ? Prove . (d) (8 pts) Now, suppose we are going to regularize the model with a ridge regularization, alllB 1 12 + 1/B 1/2) , what is the effect of increasing on bias and variance ? Show the numerical results using A value of 0.01 and 0.1