Amazon Forecast is no longer available to new customers. Existing customers of Amazon Forecast can continue to use the service as normal. [Learn more"](https://aws.amazon.com/blogs/machine-learning/transition-your-amazon-forecast-usage-to-amazon-sagemaker-canvas/)

# Prophet Algorithm
<a name="aws-forecast-recipe-prophet"></a>

[Prophet](https://facebook.github.io/prophet/) is a popular local Bayesian structural time series model. The Amazon Forecast Prophet algorithm uses the [Prophet class](https://facebook.github.io/prophet/docs/quick_start.html#python-ap) of the Python implementation of Prophet.

## How Prophet Works
<a name="aws-forecast-recipe-prophet-how-it-works"></a>

Prophet is especially useful for datasets that:
+ Contain an extended time period (months or years) of detailed historical observations (hourly, daily, or weekly)
+ Have multiple strong seasonalities
+ Include previously known important, but irregular, events
+ Have missing data points or large outliers
+ Have non-linear growth trends that are approaching a limit

Prophet is an additive regression model with a piecewise linear or logistic growth curve trend. It includes a yearly seasonal component modeled using Fourier series and a weekly seasonal component modeled using dummy variables.

For more information, see [Prophet: forecasting at scale](https://research.facebook.com/blog/2017/2/prophet-forecasting-at-scale/).

## Prophet Hyperparameters and Related Time Series
<a name="aws-forecast-recipe-prophet-hyperparamters"></a>

Amazon Forecast uses the default Prophet [hyperparameters](https://facebook.github.io/prophet/docs/quick_start.html#python-ap). Prophet also supports related time-series as features, provided to Amazon Forecast in the related time-series CSV file.