Академический Документы
Профессиональный Документы
Культура Документы
Customer
Muffins > Stuff-Mart
Expert Analysis
Using rule-based logic I have narrowed down the choice to exponential smoothing or Box-Jenkins.
I will perform an out-of-sample test to select between these two approaches.
The cumulative MAD for Exponential smoothing was 502883 and for Box-Jenkins was 435313.
The rolling out-of-sample test used a maximum horizon of 12 and generated 78 forecasts for each method.
Model Details
Warning - displayed forecasts represent adjusted values, not actual forecasts using the model shown
Expert selection
Box-Jenkins
ARIMA(0, 0, 0)*(0, 1, 1)
Term Coefficient Std. Error t-Statistic Significance
B[12] 0.85876 0.060541 14.185 1
Within-Sample Statistics
Forecast Data
Total 99135182.9439
Average 8261265.2453
Minimum 5886351.5486
Maximum 12209243.414
Overrides
Aggregated Outliers
Status Corrected
Aggregated
Date Outlier Corr. Value
2002-Feb 6471808 5952426.8793 519381.1207
2002-Sep 10150016 7435094.4953 2714921.5047
2003-Mar 9212832 11043037.9207 -1830205.9207
2004-May 4307008 6151541.951 -1844533.951
2004-Oct 9106800 8693453.4495 413346.5505
2004-Dec 12364448 12248212.0179 116235.9821
2005-Jan 8388736 7843443.8437 545292.1563
2005-Mar 8659392 8414046.9961 245345.0039
2005-Apr 5868768 5944223.2513 -75455.2513
2005-Dec 12943752 12363126.3467 580625.6533
2006-Sep 9462312 9072580.0493 389731.9507
2006-Nov 7753792 8098627.9048 -344835.9048
2006-Dec 10364640 11702637.995 -1337997.995
for each method.
he model shown
97.5 Upper
7375097.2015
7723046.1527
10974709.7844
11249769.3101
11209618.1084
13697989.0669
9555786.6227
9598334.9528
10547530.6739
8080408.5375
7782675.8428
9205164.5244
SubCategory
Muffins > Stuff-Mart > Grain
Expert Analysis
Using rule-based logic I have narrowed down the choice to exponential smoothing or Box-Jenkins.
Nonpositive series. Multiplicative seasonality ruled out.
I will perform an out-of-sample test to select between these two approaches.
The cumulative MAD for Exponential smoothing was 103331 and for Box-Jenkins was 223218.
The rolling out-of-sample test used a maximum horizon of 12 and generated 78 forecasts for each method.
Model Details
Warning - displayed forecasts represent adjusted values, not actual forecasts using the model shown
Expert selection
Exponential smoothing: No trend, Additive seasonality
NA(0.014, 0.255)
Forecast Data
Total 4484430.5262
Average 373702.5439
Minimum 79648.7209
Maximum 1105017.7255
Overrides
Aggregated Outliers
Status Corrected
Aggregated
Date Outlier Corr. Value
2002-Feb 725760 206378.8793 519381.1207
2004-Oct 1324080 910733.4495 413346.5505
2004-Dec 279936 697209.6951 -417273.6951
2005-Jan 795744 744740.4766 51003.5234
2005-Mar 388800 143454.9961 245345.0039
2005-Apr 25920 101375.2513 -75455.2513
2005-Dec 1885320 1304694.3467 580625.6533
2006-Sep 1133640 743908.0493 389731.9507
2006-Dec 1397088 1109731.6022 287356.3978
for each method.
he model shown
97.5 Upper
610475.7004
647421.7695
1141963.2099
1375247.6909
988205.9821
1610039.8742
1064426.013
674179.1043
621342.26
614179.5141
584906.6689
612591.2269
Forecast Report for COR-12-11
Expert Analysis
Using rule-based logic I have narrowed down the choice to exponential smoothing or Box-Jenkins.
I will perform an out-of-sample test to select between these two approaches.
The cumulative MAD for Exponential smoothing was 16160 and for Box-Jenkins was 17654.
The rolling out-of-sample test used a maximum horizon of 12 and generated 78 forecasts for each method.
Model Details
Expert selection
Exponential smoothing: No trend, Additive seasonality
NA(0.097, 0.460)
Within-Sample Statistics
Forecast Data
Total 1117877.8826
Average 93156.4902
Minimum 40000
Maximum 187069.804
Overrides
Outliers
Status Corrected
Expert Analysis
Using rule-based logic I have narrowed down the choice to exponential smoothing or Box-Jenkins.
Nonpositive series. Multiplicative seasonality ruled out.
I will use smoothing--the series has too many zeroes for Box-Jenkins.
Model Details
Expert selection
Exponential smoothing: No trend, Additive seasonality
NA(0.011, 0.310)
Within-Sample Statistics
Forecast Data
Total 1695378.4344
Average 141281.5362
Minimum 22080.4201
Maximum 551489.4355
Overrides
Outliers
Status Corrected
Expert Analysis
Using rule-based logic I have narrowed down the choice to exponential smoothing or Box-Jenkins.
Nonpositive series. Multiplicative seasonality ruled out.
I will use smoothing--the series has too many zeroes for Box-Jenkins.
Model Details
Expert selection
Exponential smoothing: No trend, Additive seasonality
NA(0.008, 0.291)
Within-Sample Statistics
Forecast Data
Total 1671174.2092
Average 139264.5174
Minimum 3097.8931
Maximum 411708.373
Overrides
Outliers
Status Corrected
SubCategory
Muffins > Stuff-Mart > Fruit
Expert Analysis
Using rule-based logic I have narrowed down the choice to exponential smoothing or Box-Jenkins.
I will perform an out-of-sample test to select between these two approaches.
The cumulative MAD for Exponential smoothing was 439830 and for Box-Jenkins was 363779.
The rolling out-of-sample test used a maximum horizon of 12 and generated 78 forecasts for each method.
Model Details
Warning - displayed forecasts represent adjusted values, not actual forecasts using the model shown
Expert selection
Box-Jenkins
ARIMA(1, 0, 0)*(1, 0, 1)
Term Coefficient Std. Error t-Statistic Significance
a[1] 0.29618 0.11183 2.6485 0.98979
A[12] 0.99965 0.00012263 8151.8 1
B[12] 0.87292 0.060416 14.449 1
Within-Sample Statistics
Forecast Data
Total 94650752.4177
Average 7887562.7015
Minimum 5780662.0874
Maximum 11104225.6885
Overrides
Aggregated Outliers
Status Corrected
Aggregated
Date Outlier Corr. Value
2002-Sep 9680000 6965078.4953 2714921.5047
2003-Mar 9199872 11030077.9207 -1830205.9207
2004-May 4294048 6138581.951 -1844533.951
2004-Dec 12084512 11551002.3228 533509.6772
2005-Jan 7592992 7098703.3671 494288.6329
2006-Nov 7360672 7705507.9048 -344835.9048
2006-Dec 8967552 10592906.3928 -1625354.3928
for each method.
he model shown
97.5 Upper
7315850.914
7692819.9735
10455642.8208
10497960.0273
10844941.1247
12711529.1723
9114988.0921
9547830.5245
10549910.2524
8089998.0176
7821585.3191
9216436.5894
Forecast Report for BLU-12-11
Expert Analysis
Using rule-based logic I have narrowed down the choice to exponential smoothing or Box-Jenkins.
I will perform an out-of-sample test to select between these two approaches.
The cumulative MAD for Exponential smoothing was 94643 and for Box-Jenkins was 133937.
The rolling out-of-sample test used a maximum horizon of 12 and generated 78 forecasts for each method.
Model Details
Expert selection
Additive Winters: Linear trend, Additive seasonality
LA(0.023, 0.047, 0.356)
Within-Sample Statistics
Sample size 52 No. parameters 3
Mean 1751178.75 Std. deviation 343551.66
R-square 0.81 Adj. R-square 0.8
Durbin-Watson 1.73 Ljung-Box(18) 38.6 P=1.00
Forecast error 154597.21 BIC 168189.2
MAPE 7.44 RMSE 150071.43
MAD 124040.36
Forecast Data
Total 24473245.2741
Average 2039437.1062
Minimum 1529335.8852
Maximum 2621831.9283
Overrides
Outliers
Status Corrected
Expert Analysis
Using rule-based logic I have narrowed down the choice to exponential smoothing or Box-Jenkins.
I will perform an out-of-sample test to select between these two approaches.
The cumulative MAD for Exponential smoothing was 361440 and for Box-Jenkins was 323952.
The rolling out-of-sample test used a maximum horizon of 12 and generated 78 forecasts for each method.
Model Details
Expert selection
Box-Jenkins
ARIMA(1, 0, 0)*(1, 0, 1)
Term Coefficient Std. Error t-Statistic Significance
a[1] 0.67911 0.090085 7.5385 1
A[12] 0.99881 0.006128 162.99 1
B[12] 0.86641 0.048582 17.834 1
Within-Sample Statistics
Forecast Data
Total 70177507.1436
Average 5848125.5953
Minimum 4161483.1832
Maximum 8482393.7602
Overrides
Outliers
Status Corrected