Time series forecasting with Amazon SageMaker AutoML

This blog post discusses time series forecasting using Amazon SageMaker AutoML. It covers the entire process from data preparation to model deployment, highlighting the key aspects of this methodology.

Specifically, the article covers:

• Data preparation for time series forecasting, including splitting data into train and test sets, handling timestamps, and creating datasets.
• Training a model with AutoMLV2, including defining the time series forecasting configuration and initializing the AutoML job.
• Deploying a model with AutoMLV2, identifying the best model, and creating a SageMaker model.
• Inference methods: batch inference using SageMaker Pipelines, real-time inference with SageMaker Endpoints, and asynchronous inference.
• Conclusion summarizing the benefits of using SageMaker AutoML for time series forecasting and promoting SageMaker Canvas for further exploration.