Robust Time Series Forecasting with MLOps

This solution demonstrates how to train a time series forecasting model that is robust to outliers using a Distributional TCN with Spliced Binned Pareto Distribution. We will also be covering how to deploy this solution in an endpoint. Throughout the process we will be leveraging SageMaker features that streamlines the data science process by utilizing AWS’s cloud infrastructure with the use of SageMaker Pipelines. This model can be applied to any time-series problem given that it has a sufficient amount of data.

Getting Started

1. Clone the repository on Amazon SageMaker.
2. Open the SBP_main.ipynb Jupyter Notebook.
3. Select Python 3 kernel with Pytorch 1.13 Python 3.9 CPU Optimized image.
4. Run each cell in the SBP_main.ipynb Jupyter Notebook.

Security

See CONTRIBUTING for more information.

License

This library is licensed under the MIT-0 License. See the LICENSE file.