The accompanying data file contains quarterly data on expenses (in $1,000s) over five years. For cross-validation, let the training and the validation sets comprise the periods from 2008:01 to 2015:04 and 2016:01 to 2017:04, respectively. 

a. Use the training set to implement the Holt-Winters exponential smoothing method with additive and multiplicative seasonality and compute the resulting MSE, MAD, and MAPE for the validation set. 

b. Determine the preferred model and reimplement it with the entire data set to forecast expenses for the first quarter of 2018.