The question is shown in the picture

Question 4 Assume we fit the following model: E(Y|X) = Bo + BIX1 + B2X2 + B3X3 Below are several diagnostic plots. 50 50 - 25- 25 201 Residuals Residuals -25 - -25 -50 - -50 -50 O 50 100 10.0 12.5 15.0 17.5 Fitted X1 50 - 50 25 25 Residuals Residuals -25 -25 -50 -50 - 2 A 12 14 16 18 X2 X3 I claim that from these plots, I see evidence of violations of the assumptions that E(c|X) = 0 and var(|X) is constant. Answer the following questions. a) Which variable(s) would you start looking at first to try and address the zero mean error assumption violation? Justify your answer. b) Which variable(s) would you start looking at first to try and address the homogeneity of variance assumption violation? Justify your answer. c) Which variable(s) would you start looking at first to try and address the independence assumption violation? Justify your answer. d) Given your answers for the previous parts, make a suggestion of a new model to fit that hopefully addresses the assumption violations