Time-series Forecasting Evaluation and Deployment (Time-FED)

About this technology

Overview

The Time-FED system is a machine learning system designed for preparing time-series for a variety of ML models. At its core, it is a sophisticated software framework for data scientists to transform time-series datasets into ML-ready datasets and trained models. The produced models are then ready for a real-time environment for forecasting purposes. One of the key features of Time-FED is its configuration flexibility for ML research. The system is designed by and for data science. By taking a generalized approach, Time-FED is capable of adapting to new use-cases all via its vast configuration. This framework allows for new infusions and use-case opportunities, making it an attractive collaboration platform for researchers interested in ML software developed by JPL.