4. Considering our forecasting handout and methods computed, a. What other methods could be used to forecast the demand for November? b. Which method would you recommend and why? FORECASTING Naive 3 month moving avg. Forecast Error Weighed moving avg. Forecast Error Exponential a = .3 Forecast Error Forecast Error Month Orders Jan 120 Feb 90 March 100 April 75 May 110 June 50 July 75 August 130 Sept 110 Oct 90 Nov Dec 120 90 100 75 110 103.3 88.3 100.1 85.8 96.75 120 111 107.7 97.89 101.52 95 Naive Forecast: F =D , Simple Moving Average: MA n 3 month moving average the first month you can predict is 5 month moving average the first month you can predict is Longer period moving average reacts Longer period moving average reacts Weighted Moving Average: WMA, - WDwhere w. = 1.00 wD (This example uses the weights of 5 for the most recent month, 3 two months ago, and. 17 three months ago) Exponential Smoothing: F-1 = F,+ a (D-Fi) or F1 = a D. + (1 - a) F Algebraic translation Ft+1 = F,+ a (D. - F.) multiply out the parenthesis F1+1 = F,+ a D-a F isolate the F F-1 = a D,+ (F - a F) remove the F, from the parenthesis F1 = a D + (1 - a) F an easier equation to drop values into a varies between 0 and 1 and reflects the amount of error incorporated into the next forecast So, if a = 0, F- = if a = 1, F1 = Start exponential smoothing forecasts with a nave forecast ERROR = ACTUAL DEMAND - FORECAST Bias measured by Cumulative Forecast Error = Le OR Mean Forecast Error (MFE) = xe Magnitude measured by Mean Absolute Deviation (MAD) =( ED-F.D