Amazon Forecast
Developer Guide

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

The Amazon Forecast API will undergo significant changes during scheduled maintenance occurring from 10 AM on 7/22/19 until 10 AM on 7/23/19. During maintenance, access to the Forecast APIs and console might be interrupted.

After 7/22/19, your Forecast resources (datasets, predictors, and forecasts) will no longer be available. However, you can save your forecasts for future use. We recommend using the CreateForecastExportJob API to save them to your S3 bucket before 7/22/19.

After maintenance concludes, before using the APIs, you must download a new SDK and modify your existing code to reflect the syntax changes. If you use only the console, you won’t need to make any changes.

We will provide new API documentation before scheduled maintenance begins. If you have questions, contact

Predictors

To generate a forecast, you train a predictor, and then deploy it. After the predictor is trained, you can get accuracy metrics for that predictor. You evaluate the metrics and decide whether to deploy the predictor to generate a forecast.

Creating Predictors

Amazon Forecast trains forecasting models called predictors. To create a predictor, you use the CreatePredictor operation. To manage predictors, you can use the Amazon Forecast console, the AWS Command Line Interface (AWS CLI), or the AWS SDKs.

To create a predictor, you provide the following:

  • A dataset group – The dataset group provides data for training the predictor. For more information, see Datasets.

  • A recipe – A recipe is a combination of an algorithm and the featurization required to train a predictor. Amazon Forecast provides a set of predefined recipes. If you don't know which recipe to choose, use the PerformAutoML option. This option tells Amazon Forecast to evaluate all recipes and choose the best configuration for your datasets. With this option, predictor training can take longer, but you don't need to worry about choosing the right recipe. For more information, see Choosing an Amazon Forecast Recipe.

Each predictor has a name and a version ID.

If you update your datasets, you might want to retrain your predictors. You can retrain predictors by setting a schedule in your initial call to the CreatePredictor or RetrainPredictor operations. To update the predictor configuration, use the UpdatePredictor operation.

After creating an Amazon Forecast predictor, review the metrics to see how well the predictor performed on your data during the evaluation process.

Predictor Evaluation

After you create a predictor, you can evaluate the accuracy of the forecasts it generates by running the GetAccuracyMetrics operation using the console or the AWS CLI. Always evaluate metrics before deciding whether to use the predictor to generate forecasts.

How It Works: Next Topic

Forecasts