Given the rapid growth of available methods, selecting the right time series forecasting model is increasingly complex.
Benchmarking frameworks address this by enabling systematic, reproducible comparisons across models and datasets. This tutorial introduces the sktime benchmarking module, which allows users to evaluate any compatible forecaster with minimal code. We will demonstrate how to apply the module to real datasets, design benchmarking experiments, and interpret the results to guide model selection.
You will:
- Get to know sktime and its time series forecasting and analysis capabilities.
- Learn how to use the sktime benchmarking module to set up, run, and analyze forecasting benchmarks on custom datasets.
- Learn about the roadmap of the benchmarking module and how to contribute to its ongoing development.
Read a detailed description here on GitHub.
