Problem 3. In this problem (and the next one), you will use the 'Baseball Salary Data' found under the item \"Baseball Salary Data and Some Related Files\" in Datasets at eCampus. The item includes the data (a .csv le), a .txt le with some background information on the data and the variables, and a .pdf le with some basic R commands to get you started with the analysis of the data, including how to: (a) read the data into R, and (b) use the command lm(-) when you have a large number of predictors. Further, under R Codes and Instructions, the item \"Some R Commands for the Baseball Salary Data\" enlists several R codes and summary les (.pdf) regarding analysis of the 'Baseball Salary Data\" which you may nd useful. Make sure to read the descriptions given under each of these items! Beware that any R code given in eCampus uses a log transformation of the \"salary\" variable. But here, you are not going to use the log transformation! Using the 'Baseball Salary Data', do the following: (a) Fit a linear regression model with \"salary\" as the response (and not log(salary), as done in class/lab sessions) and the other 16 variables (excluding \"names\") as the predictors. (b) What percentage of the variation in salaries is explained by the linear model above? (c) Comment on the coeicient of the predictor \"hits\". Is this coeicient consistent with what your intuition says should be the relationship between number of hits and salary? Why or why not? ((1) Test the null hypothesis (using level of signicance 0: = 0.05) that none of the 16 predictors is related to salary. What is the proper conclusion about the linear model above and its utility? (e) Test the null hypothesis (using level of signicance or = 0.05) that the variables \"batting average\" , \"on base percentage\