For exponential smoothing models, the (alpha) parameter changes :

1. the accuracy of the model. A larger alpha is more accurate for smaller datasets while a smaller alpha is more accurate for larger datasets.
2. the accuracy of the model. A larger alpha is more accurate for larger datasets while a smaller alpha is more accurate for smaller datasets.
3. the responsiveness of the model. A larger alpha emphasises the long term trend, while a smaller alpha emphasises the most recent data.
4. the responsiveness of the model. A larger alpha emphasises the most recent data, while a smaller alpha emphasises the long term trend.