Online shopping giant Amazon has released a machine learning forecasting tool that has been developed using the same systems that run its shopping platform Amazon.com
Amazon forecast uses machine learning to create custom models that developers can use to make predictions about a host of retail or industrial variables. Forecast takes into account factors such as product features, seasonality and store locations, it then uses machine learning algorithms to understand the complex patterns that can arise when these intersect.
Forecast combines time series data and business dependant variables to build a forecast of future needs. If for example you are selling a particular type of product like cloths or food, if can be difficult to judge how seasons and customer interest will vary over the year. Amazon forecast uses historical data to predicate how variables like seasons will affect product sales, thus allowing firms to get a better handled over stock control.
The machine learning application of Forecast can be applied to a host of industries. CJ Logistics, one of the leading logistic providers in Korea, uses the service to help reduce the cost of warehouse space.
YoungSoo Kim, Vice President of TES Strategy Unit for CJ Logistics commented in a release that: “Amazon Forecast has been applied to CJ Logistics’ parcel volume forecasting process to optimize the amount of human resources, transportation, and warehouse space we provision to meet demand. Amazon Forecast allows us to use sophisticated machine learning-based forecasting techniques without building our own system. We have a clear method for increasing our operational efficiency by using Amazon Forecast.”
Amazon Forecast
Developers do not need to be well versed in machine learning programming to use the tool, as Amazon Forecast will automatically establish a data pipeline that ingests data from your chosen sources which it then uses to train a machine learning model. This model can be used to glean detailed metrics and produce a forecast in relation to the chosen business outcome.
The machine learning interface has been created with an easy to use API so developers of varying skills can use the service in AWS. The simplified API lets developers create machine learning models in less than five API calls.
Swami Sivasubramanian, Vice President of Amazon Machine Learning commented that: “Amazon Forecast now offers the forecasting expertise from Amazon’s first 25 years of building the world’s largest ecommerce business in a managed service for any company to leverage.”
“We’ve built sophisticated, machine learning forecasting algorithms over many years that our customers can now use in Amazon Forecast without having to know anything about machine learning themselves. We can’t wait to see how our customers use the service to reduce operating expenses and inefficiencies, ensure higher resource and product availability, deliver products faster, and lower costs to delight their customers.”