1. Use the 36 successive values of demand data given in problem 2 to determine the number of periods to include in a moving average so that the MAD is minimized. You can construct a spreadsheet with columns for month number, demand, sum of n months, forecast, deviation, and absolute deviation. Sum the absolute deviation column for months 13 through 36 to get the numerator for the MAD. Use 450 for the initial forecast, and use a 2-month moving average. Record the MAD, then repeat with a 3-month moving average, etc., until you find the number of months in a moving average that minimizes the MAD. Use the prior month (a l-month average) as a forecast. Which forecast model seems to fit the data best? Why?