Continuing with the DataNet forecasting problems, develop a double exponential smoothing model using smoothing constants a = 0.20 and b = 0.20. As starting values, use the least squares trend line slope and intercept values.
a. Compute the MAD for this model.
b. Plot the forecast values against the actual data.
c. Compare this with a linear trend model. Which forecast method would you use? Explain your rationale.
d. Use the same starting values but try different smoothing constants say, [(α, β) = (0.10, 0.30), (0.15, 0.25), and (0.30, 0.10) in an effort to reduce the MAD value. Prepare a short report that summarizes your efforts.