Which one is the best model for sales data and why?
Method |
MAPE |
MAD |
MSE |
|
Naïve |
12,5 |
1174,1 |
3910351 |
|
Naïve-Seasonal |
4,5 |
359,9 |
219893 |
|
Moving Average |
k = 12 |
11 |
1011 |
2588085 |
Single Exponential |
Level = 0.1 |
11 |
1049 |
2770820 |
Holt’s |
Level = 0.1 Trend = 0.1 |
12 |
1123 |
2862244 |
Additive Winter’s |
Level = 0.3 Trend = 0.1 Seasonal = 0.5 |
2,7 |
223,2 |
93724,5 |
Multiplicative Winter’s |
Level = 0.4 Trend = 0.1 Seasonal = 0.5 |
2,2 |
186,7 |
72648,6 |
Additive( trend plus seasonal ) |
6 |
472 |
359304 |
|
Multiplicative( trend plus seasonal) |
6 |
474 |
364564 |
Additive Winter's is the best model.
Because it has the lowest MAPE, MAD & MSE.
MAPE, MAD & MSE are the measure of prediction accuracy of a forecasting method. Lower the values, better model.
Which one is the best model for sales data and why? Method MAPE MAD MSE Naïve...