Consider the various forecasting methods in Exercise 2.6. Compute their forecasting errors using MAD and BIAS. Based on the forecast errors, which method would you recommend and why?
Exercise 2.6
The monthly sales data for the year 2011 for Datastream, Inc., which produces network routers for small companies, are given in Table 2.18. The sales manager wants a forecast of sales for the first quarter of year 2012 (months 13-15). Compute the forecasts using the following methods:
(a) Last value method
(b) Averaging method
(c) Three-month moving average method
(d) Exponential smoothing with α = 0.25
(e) Holt's method with α = 0.4 and β = 0.5
(f) Given the sales pattern for 2011, do any of these methods seem inappropriate for obtaining the forecasts? Is there another method you would recommend to get a better forecast for months 13, 14, and 15.
