New framework for distributional regression using ReLU neural networks

This paper introduces a novel framework for distributional regression that accommodates arbitrary response variables. It models conditional event probabilities using measurable sets, extending the approach to general response spaces. The research proposes two estimators, one based on separate regression of event indicators and another that enforces inclusion-consistent predictions via a directed acyclic graph. Error bounds are established for both, with specialized convergence rates for dense ReLU networks that bypass the curse of dimensionality. The framework's practical utility is validated through simulations and real-world experiments. AI

IMPACT Introduces a new statistical framework for machine learning that could improve the accuracy and applicability of regression models.

RANK_REASON The item is an academic paper detailing a new statistical framework and estimators for machine learning. [lever_c_demoted from research: ic=1 ai=1.0]

Read on arXiv stat.ML →

• alphaXiv
• arXiv
• CatalyzeX Code Finder for Papers
• DagsHub
• directed acyclic graph
• Gotit.pub
• Hugging Face
• rectifier
• ScienceCast

• paper
• other

AI-generated summary · Google Gemini · from 1 sources. How we write summaries →