Nixtla

StatsForecast 2.0.3

A baseline for comparing forecasts

Open in Studio

What to provide

Enter 12 to 512 consecutive, evenly spaced observations and choose a horizon of 1 to 24 periods.

Example input

Series: 12, 15, 17, 16, 19, 18, 20, 24, 22, 25, 26, 29
Horizon: 6

How it works

Studio holds out the latest observations and compares Naive and nonseasonal AutoETS by mean absolute error. It returns the better method’s forecast and an 80% interval.

How to read the result

Use it as a demand forecasting baseline. Naive repeats the last value. AutoETS can model level and trend. The comparison uses a single validation window.

Credits and licenses

StatsForecast 2.0.3 · NixtlaApache 2.0
Official documentation