/Fedot-TS-Benchmark

Benchmarking Fedot on timeseries forecasting tasks

Primary LanguageJupyter Notebook

FEDOT Timeseries benchmarking [OBSOLETE, see pytsbe for actual examples]

The repository contains some experiments and benchmarks of our AutoML framework (FEDOT) on timeseries forecasting tasks. We are planning to compare state-of-the-art AutoML frameworks in this task and design the suitable strategies for automatic data-driven model identification that will be integrated into Fedot.

This is an ongoing research.

Current progress

The proof-of-concept of AutoML for timeseries forecasting was already implemented in FEDOT and some promising results were obtained. For instance, check our paper "Automated Data-Driven Approach for Gap Filling in the Time Series Using Evolutionary Learning" about timeseries gap filling with FEDOT.

Supported by