Data set is already in the r studio.

Problem 2 The cheddar data from the faraway library has data on samples of cheddar cheese. For each of 30 samples, a subjective taste score is given along with checmical composition measurements labeled Acetic, H2S, and Lactic. a) Fit a multiple linear regression model for predicting taste from the three chemical composition measurements. Which of the variables have statistically significant coefficients at level 0.05? Answer: b) Plot studentized residuals versus fitted values for the model in a), including a horizontal line at height 0 for reference. Is there any evidence of outliers, heteroscedasticity or curvature in the plot? Explain briefly. Answer: c) Use the Box-Cox method to find an optimal transformation of the response. What is the optimal power A of the response, based on the log-likelihood? Answer: d) Does the 95% confidence interval for the Box-Cox transformation parameter A include A = 1? What is your interpretation of this? Answer: e) Use the optimal transformation found in c) to refit the data with the transformed response depending on the three chemical composition measurements. Plot studentized residuals versus fitted values for this transformed response model. Compare the graph qualitatively to the graph in Part b). Note: you can include a transformation of a variable in a formula by enclosing the transformation inside the function I( ), as in Im(I(y 3)~x, . . .)