Overview

Probabilistic forecasting estimates ranges, quantiles, event probabilities, or full predictive distributions rather than only a single future value. Methods include Bayesian and parametric models, ensemble post-processing and conformal calibration, and neural generative approaches such as diffusion and flow matching; language-model agents also forecast real-world events using evidence and probability aggregation. Researchers evaluate forecasts with proper scores such as CRPS and Brier score, alongside calibration, sharpness, and benchmarks in weather, energy, health, finance, and human mobility. A central challenge is balancing accuracy, reliable uncertainty, computational cost, and usefulness for downstream decisions, since strong average scores may hide failures in extremes or dependencies between outcomes.

Latest papers

CardsList

245 papers19 papers in the last four weeks, down 58% on the four weeks before. 0.1% of all new papers.

  1. CRPS-LAM: Probabilistic Regional Weather Forecasting with Continuous Ranked Probability Score

    Oct 10, 2025Erik Larsson, Joel Oskarsson, Tomas Landelius +1Graph Neural NetworksEnsemble Learning

  2. Huracan: A skillful end-to-end data-driven system for ensemble data assimilation and weather prediction

    Aug 25, 2025Zekun Ni, Jonathan Weyn, Hang Zhang +7Ensemble LearningData Assimilation

  3. Edge Selection for the Effective use of Piecewise-Constant Distributions as Neural Network Outputs for Event Prediction

    Jul 29, 2025Kevin Doran, Tom BadenEvent Time PredictionProbabilistic Forecasting

  4. Isotonic Quantile Regression Averaging for uncertainty quantification of electricity price forecasts

    Jul 20, 2025Arkadiusz Lipiecki, Bartosz UniejewskiProbabilistic ForecastingQuantile Regression

  5. A-UTE: Advection Informed, Uncertainty Aware Temperature Emulator

    Oct 29, 2024Hira Saleem, Cormac Purcell, Steven Sherwood +1Climate EmulationNeural Surrogate Modeling