Reproduce “Ensemble Modeling for Time Series Forecasting: an Adaptive Robust Optimization Approach” on OpenResearch
Turn your favorite coding agents into research agents that review literature, develop hypotheses, and run experiments.
Universal app for Apple silicon and Intel. Requires macOS 11 or later.
About this paper
Dimitris Bertsimas, Leonard Boussioux
An Adaptive Robust Optimization (ARO) framework creates robust time series ensembles by dynamically adjusting model weights based on past errors. This method demonstrates superior accuracy and protection against worst-case errors compared to traditional approaches in various forecasting scenarios.
Key insights
- Adaptive Robust Optimization can be effectively applied to dynamic machine learning problems, specifically for time series ensemble forecasting.
- A complex min-max ARO problem for ensemble weighting can be equivalently transformed into a tractable regularized regression formulation.
- Allowing ensemble weights to adapt based on recent forecast errors significantly enhances robustness and performance, especially when models experience drift.
Summary by alphaXiv.
Tip: change arxiv.org to autoarxiv.org in any arXiv abstract link to open the paper here.