Companies today struggle to accurately forecast demand using existing tools that fail to combine historical time series data with relevant independent variables. Based on the same technology used at Amazon.com, Amazon Forecast uses machine learning to combine historical time series data with additional variables to build more accurate forecasts. In this tech talk, learn how to build and deploy models, and measure business impact using Amazon Forecast. We’ll walk through all of the steps, cover best practices, and answer common questions to help you get started.
This technical paper introduces forecasting, its terminology, challenges, and use cases. This document uses a case study to reinforce forecasting concepts, forecasting steps, and references how Amazon Forecast can help solve the many practical challenges in real-world forecasting problems.
Provides a conceptual overview of Amazon Forecast, includes detailed instructions for using the various features, and provides a complete API reference for developers.
You can find additional samples to get started with on GitHub.
Solution implementation example
The Improving Your Forecast with Machine Learning solution enables customers to bring models to production faster and with less overhead costs by generating, testing, comparing, and iterating on forecasts from Amazon Forecast. You can use this solution to accurately predict retail inventory demand, supply-chain planning, workforce status, web traffic forecasting, and more.