This data set shows the bikers on Brooklyn Bridge 7 days of the week compared to the number on all bridges to give us the rate that bikers travel the Brooklyn bridge.
Here we create a forecasting table to show the data over a specific amount of time in four different methods.
| pred.mv | pred.naive | pred.snaive | pred.rwf |
|---|---|---|---|
| 0.1468914 | 0.1352921 | 0.1352921 | 0.1344407 |
| 0.1468914 | 0.1352921 | 0.1352921 | 0.1335893 |
| 0.1468914 | 0.1352921 | 0.1352921 | 0.1327379 |
For this forecasting table we can see that the nativ and snative methods are both similar and rfw the best method to use for forecasting.
We can see that Naive and Snaive are both on top of eachother and that Moving Value is closet to the actual point that is graphed showing that they are in fact the closest predicted rate for values for rate
Intuitively, moving average should produce non-constant forecast values. What
D. Accuracy Metrics
We will use the mean absolute prediction error (MAPE) to compare the performance of the four forecasting methods.
| MAPE | MAD | MSE | |
|---|---|---|---|
| Moving Average | 7.558001 | 0.0384566 | 0.0003419 |
| Naive | 11.713788 | 0.0588914 | 0.0006627 |
| Seasonal Naive | 11.713788 | 0.0588914 | 0.0006627 |
| Drift | 12.815859 | 0.0639998 | 0.0007367 |
In summary, the moving average method has the best performance because it has the lowest predicted error out of all of the methods.