Amazon Forecast now supports generating predictions for 10X more items

Posted on: Jul 7, 2020

Amazon Forecast is a fully managed service that uses machine learning (ML) to generate accurate forecasts without requiring any prior ML experience. Amazon Forecast can be used in a wide variety of use cases, including product demand forecasting, inventory planning, workforce planning and cloud infrastructure usage forecasting.

Today we are excited to announce that Amazon Forecast can now generate predictions for 10X more items. Until today, customers were limited to generating forecasts for one hundred thousand unique items based on the prior limits imposed by the CreatePredictor API. In some use cases, customers with a large number of items have either had to limit their model to a subset of items or generate models that are specific to certain items/categories. Both approaches require additional data preparation and may not fully leverage the inherent advantages of training deep learning algorithms with large data sets. Today, we are raising that limit to one million unique items for data frequencies of daily and above. Additionally, we have also increased the number of historical observations that can be supported in the target time series dataset from hundred million to one billion. This allows customers to include a maximum of one thousand observations per item (when having one million items) providing sufficient historical data when generating models.

The limit increase is now available in US East (N. Virginia, Ohio), US West (Oregon), Europe (Ireland, Frankfurt),and Asia Pacific (Tokyo, Singapore, Seoul, Sydney, Mumbai). It can be leveraged when using the CreatePredictor or CreateForecast APIs. Please visit the Amazon Forecast getting started guide or updated service limits for more information.