Parametric Prior Mapping Framework for Non-stationary Probabilistic Time Series Forecasting
The article introduces a new framework called Parametric Prior Mapping (PPM) for forecasting non-stationary probabilistic time series. PPM combines the strengths of parametric methods and deep generative models to improve predictive accuracy and efficiency. Empirical results indicate that PPM outperforms existing methods in handling non-stationary data.
- ▪Parametric Prior Mapping (PPM) injects parametric structural priors into generative modeling processes.
- ▪The framework allows for dynamic, adaptive priors that enhance the learning of complex predictive distributions.
- ▪PPM has been shown to yield precise forecasts with well-calibrated uncertainty estimates.
Opening excerpt (first ~120 words) tap to expand
Computer Science > Machine Learning arXiv:2605.23402 (cs) [Submitted on 22 May 2026] Title:Parametric Prior Mapping Framework for Non-stationary Probabilistic Time Series Forecasting Authors:Jinglin Li, Jun Tan, QI Fang, Ning Gui View a PDF of the paper titled Parametric Prior Mapping Framework for Non-stationary Probabilistic Time Series Forecasting, by Jinglin Li and 2 other authors View PDF HTML (experimental) Abstract:Effectively modeling non-stationary dynamics in probabilistic multivariate time series(MTS) forecasting requires balancing expressiveness with robustness. Existing parametric approaches benefit from strong inductive biases but lack flexibility, whereas deep generative models struggle to capture complex temporal dependencies without extensive data and computation.
…
Excerpt limited to ~120 words for fair-use compliance. The full article is at arXiv cs.AI.
Discussion
0 commentsMore from arXiv cs.AI
