Please solve all the parts of excercise number 1 to me

Please solve all the parts of excercise number 1 to me

Imagine a dataset containing the following variables: You are asked to construct a regression model explaining the influence of the independent variables temp, age, income, education and drug on the dependent variable death. a) Why would it be inappropriate to employ a linear regression model of the form yi=xiT+i,iiidN(0,2) ? Give a succinct answer (one sentence at most). b) You have chosen to apply a logistic regression model (Logit Model). Show that the response function h(i)=1+expiexpi is stricly monotonically increasing and maps to [0,1]. Now assume that you want to regress income on an arbitrary number of explanatory variables available to you in a separate dataset. You choose to do so with a linear regression model. You consider different transformations of the data as well as differing choices of explanatory variables and calculate the following model selection criteria for your two favourite candidate models: c) Which model would you choose based on the calculated model selection criteria? Give a succinct (at most two sentences) explanation based on an evaluation of all three given criteria. d) Is it more reasonable to rely on Radj2 or R2? e) Will AIC and Radj2 always agree on the model to be chosen? 1 f) Is it possible for R2 or Radj2 to assume negative values? g) Assume that model 1 and 2 are both based on 200 observations. How many parameters does each model have based on the table above? Since the values in the table have been rounded, you will have to round your answer in turn