Question: We want to decide on the best forecasting technique that suits the trend in our gasoline sales. For that reason we recorded the actual weekly sales of gasoline for the past 12 weeks. We want to apply different forecasting techniques and then compare the resulting forecasts to the actual sales presented in table 3 below.

Calculate the forecasted sales using the following techniques. In order to evaluate the results calculate the mean absolute deviation (MAD) and the mean squared error (MSE) for each technique.

(a) 3-period moving average

(b) Weighted moving average. The most recent observation has a weight three times as great as that given to the oldest observation and the next oldest observation has a weight twice as great as the oldest.

(c) Exponential smoothing using a = 0.2. Since in this method the forecast of a period is a function of both the forecast and the actual values of the last period, start by assuming that F2 = A1 = 17.

(d) What is the best forecasting technique for the above data? Why?