Skip to main content

Amazon Forecast

What is it​

A fully managed service that uses statistical and machine learning algorithms to provide highly accurate time series forecasts.

What it's for​

Automate the data forecasting process, allowing you to predict business outcomes with ease and accuracy, such as product demand, sales, network traffic, and resource consumption.

Use cases​

  • Product demand forecasting for inventory optimization
  • Sales forecasting for financial and marketing planning
  • Network traffic forecasting for capacity planning
  • Energy consumption forecasting for cost optimization
  • Financial results forecasting for strategic planning

Key points​

  • Fully managed: No servers to provision or manage
  • Machine Learning: Uses the same ML technology used at Amazon.com for forecasting
  • Advanced algorithms: Supports a variety of forecasting algorithms, including neural networks and statistical models
  • Automation: Automates much of the forecasting process, from data preparation to forecast generation
  • Accuracy: Designed to provide highly accurate forecasts, even with complex data
  • Integration: Integrates with Amazon S3 for data input and output

Comparison with traditional forecasting methods:​

  • Amazon Forecast: Offers a fully managed ML-based forecasting service that automates the process and provides more accurate forecasts, especially for large volumes of data and complex patterns. Reduces the need for data science expertise and infrastructure.
  • Traditional forecasting methods (e.g., spreadsheets, simple statistical models): Can be time-consuming, less accurate for complex data, and require specialized knowledge in statistics and modeling. Don't scale well for large volumes of data.