[ad_1]
Amazon's AWS today launched Amazon Forecast, a new predefined machine learning tool that will make it easier for developers to generate predictions based on time series data. Although predictions are the most standard use case for machine learning, their construction still requires some skill. Of course, Amazon has already built many of these models for its own needs. It converts them now essentially into a product.
"With just three clicks, you can provide us with information and forecasts," said Andy Jassy, CEO of AWS, in the AWS re: Invent speech. "It's very simple and when we compare our customers with the private beta and with ourselves, we provide more accurate forecasts up to 50% more accurate than what people used to do, at one-tenth the cost of traditional supply chain software.
Amazon, in its retail sector, has built a number of models to manage its own data. It's essentially the same technology that Amazon uses to forecast demand on its retail site. Users provide the company with all the data in their supply chain, and then provide the service with variables that may affect the forecast.
In the background, AWS examines the data and the signal, then chooses from among eight different predefined algorithms, shapes the model, refines it, and provides the predictions.
AWS also facilitates the integration of this service with SAP and Oracle supply chain tools, as well as with Amazon's new Timestream database service.
The service is not necessarily cheap, but it will certainly allow developers to save a lot of time.
Source link
