Using the Boston dataset:

Questions

• Use the sample() function to create a train set and store it in train (sample 300 observations out of 506)
• Create a linear regression model lm.fit1 with medv as the response and lstat as the predictor
• Now create an adaption lm.fit2 where you add a second order term of lstat
• Now create an adaption lm.fit3 where you add a second order and third order term of lstat

• For each model calculate the validation \(MSE\) and store them in mse1, mse2 and mse3 respectively

• MC1: Which model performed the best based on the validation \(MSE\)? (We see that different versions of R and different operating systems can have an impact on the answer of this question. Therefore, base you answer on the output you receive on Dodona after you submitted your code. This will be consistent for everyone)
  • 1: lm.fit1
  • 2: lm.fit2
  • 3: lm.fit3

Assume that:

• The MASS library has been loaded
• The Boston dataset has been loaded and attached
• The seed has been set on 1 (do not overwrite the seed yourself)