The following fits a gam model to data that have a strong sequential correlation (see Section 9.1 for the basic time series concepts assumed in this exercise): (a) Run the code several times. (Be...

The following fits a gam model to data that have a strong sequential correlation (see Section 9.1 for the basic time series concepts assumed in this exercise):

(a) Run the code several times. (Be sure, on each occasion, to simulate a new data frame xy.) Is the function gam() overfitting? What is overfitting in this context? Compare with the result from re-running the code with ar=0.

(b) Repeat, now with ar=-0.75 in the code that generates the sequentially correlated series. Why is the result so very different?