AWS Taps Machine Learning for New Forecast Service

Amazon Web Services Inc. (AWS) announced the general availability of Amazon Forecast, which uses machine learning to fine-tune its prognostication capabilities.

Amazon Forecast doesn't require users to have machine learning skills of their own, however, as they need only input historical time-series data and other associated data such as variables that might affect predictions. This metadata that's taken into account along with historical time-series data can range from store sizes to a widget color choice. Time-series data, meanwhile, can range from weekly product sales to daily inventory levels to hourly Web site visits and more.

While a natural fit for use cases such as retail sales forecasting, for example, AWS said its Forecast service also lends itself to logistics, finance, advertising and so on.

Business predictions are reportedly up to 50 percent more accurate when time-series data is combined with that additional metadata that provides additional variables for machine learning algorithms to crunch.

The machine learning functionality provides an edge over existing tools used to forecast business trends, which run the gamut from single spreadsheets to complex enterprise software.

Amazon Forecast
[Click on image for larger view.] Amazon Forecast (source: AWS)

"These tools build forecasts by looking at a historical series of data, which is called time series data," AWS said. "For example, such tools may try to predict the future sales of a raincoat by looking only at its previous sales data with the underlying assumption that the future is determined by the past. This approach can struggle to produce accurate forecasts for large sets of data that have irregular trends. Also, it fails to easily combine data series that change over time (such as price, discounts, Web traffic, and number of employees) with relevant independent variables like product features and store locations."

The fully managed service, which has been in preview since being unveiled at the re:Invent 2018 conference last November, is available in the Northern Virginia, Ohio, Oregon, Tokyo, Singapore and Ireland AWS regions.

About the Author

David Ramel is an editor and writer for Converge360.


Subscribe on YouTube