logo
CanyamAI summaries for academic research
language
MobileMobile
  • Phone FeatureWeChat Mini Program
EnglishEnglish
noticeNotice
BlogBlog
delete
arrow
Return

Dynamical errors in machine learning forecasts

delete2025-10-15
delete0
PRE
AI
F
Fang Zhou
G
Gianmarco Mengaldo *
DOI：10.1016/j.chaos.2025.117376delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
• We assess ML forecasts using local complexity and persistence to test dynamical consistency. • ML forecast errors (MSE, MAE) correlate with local dynamical complexity and persistence. • We identify three ML error regimes that may enable a priori, dynamics-guided training.

Journal

C
chaos, solitons & fractals
IF:
0
Papers:
851
Citations:
1

Organization

N
National University of Singapore
Scholars:
7.5W
Papers: 6.5W
Citations: 11.4W