Problem 1. [3 marks] Consider tting the simple linear regression model y 2 BO + 6133 + 5 to a dataset (mg-,1\") for i = 1, 2, . . . .71. Let 1}.- be the resulting tted value and ez- be the residual. (1a) [1 mark] When should we consider transforming y and t the linear model '9" = 50 + 5153 + 5' to (351-, yg) where y; denotes the transformed 11,-? Your answer must be based on the plot of (122-, 81'). (1b) [1 mark] If y is Poisson or count data, which transformation is appropriate? (1c) [1 mark] Name one or two ways for choosing a transformation. No need to explain