here in the orange cell ssion on "deseasonalized values" as Y range and "Period" as X range. The output is below. "Period" is the slope. s is "slope =0". The p-value for slope is 4.018E-7 (= 0.0000004018) <5% significance level, so we reject e reject slope beign 0, which means that most likely we have a non-zero slope showing a linear trend ecast for day 60 = intercept + slope * 60 (the computed cell below has a formula for calculation) = seasonality index of 90.58% * deseasonalized forecast deseasonalized demand 384.18841 371.41049 385.56732 397.909 sonalized forecast 391.55665 me forecast as of day 57 because we don't have any new information from that time until day 60 ay 53 as forecast for day 54, then obtain the forecast for day 55 using the updating formula. The numbe hing smoothing constant 0.7 384.18841 375.24386 382.47028 393.27739 me forecast as of day 57 because we don't have any new information from that time until day 60 = seasonality index of 90.58% * deseasonalized forecast 354.67419
Upper 95% Lower 95.0%Upper 95.0% 358.07933 344.08649 358.07933 0.8277921 0.4007166 0.8277921
t the null. ers below have formulas which show the process of ca
INPUTS Number of Periods of Data Collected = Performance measures for forecasted values below MSE = MAD = MAPE = LAD = Absolute Error Absolute Period Value Forecasted values Error Error Squared % Error
Smoothing constant MSE = MAD = MAPE = LAD = MSE = MAD = MAPE = LAD = MSE = MAD = MAPE = LAD = MSE = 151.32 MAD = 10.39 MAPE = 2.89% LAD = 27.61 Performance measures for Exponential Smoothing Method Performance measures for Regression Method Performance measures for Holt's Method (from forecast in sheet "Holt's") Performance measures for Classical Decomposition Method (from "Class. Decomp." worksheet)
You've reached the end of your free preview.
Want to read all 22 pages?
- Winter '10
- 5%, Holt, 95.0%, 2.89%, 53 54 55 56 3-day, 60 90.58%