Amazon Forecast
Developer Guide

This is prerelease documentation for a service in preview release. It is subject to change.

Exponential Smoothing (ETS) Recipe

The Amazon Forecast Exponential Smoothing (ETS) recipe is compatible with the open-source implementation.

How ETS Works

The Forecast Exponential Smoothing (ETS) recipe uses the exponential window function to assign exponentially decreasing weights over time, instead of constant weights as applied in the simple moving average techniques, for smoothing the input time series dataset. You can apply it to datasets with seasonality and some other prior assumptions about the data.

ETS Hyperparameters

For information about the hyperparameters used for ETS, see the ets function of the forecast package of the Comprehensive R Archive Network.

Tune ETS Models

For information about how to tune an ETS model, see the ets function of the forecast package of the Comprehensive R Archive Network.