DeepL'evy: Learning Heavy-Tailed Uncertainty in Highly Volatile Time Series

Modeling uncertainty in heavy-tailed time series remains a critical challenge for deep probabilistic forecasting models, which often struggle to capture abrupt, extreme events. While L'evy stable distributions offer a natural framework for modeling such non-Gaussian behaviors, the intractability of their probability density functions severely limits conventional likelihood-based inference. To address this, we introduce DeepL'evy, a neural framework that learns mixtures of L'evy stable distributions by minimizing the discrepancy between empirical and parametric characteristic functions. DeepL'evy incorporates a mixture mechanism that adaptively learns context-dependent weights and parameters over multiple L'evy components, enabling flexible multi-horizon uncertainty modeling. Evaluations on both real and synthetic datasets demonstrate that DeepL'evy outperforms state-of-the-art deep probabilistic forecasting approaches in tail risk metrics, especially under extreme volatility.