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

A unified algebraic framework for classification performance evaluation

delete2026-09-09
delete0
PRE
AI
R
Ronaldo C. Prati
DOI：10.1016/j.inffus.2026.104767delete
deleteOriginal
deleteOriginal request for help
deleteShare
deleteSave
Abstract

Abstract

En 中文
• Indicator matrices and aggregation operators unify classification evaluation. • Any binary measure extends to multiclass, multilabel, ordinal and cost settings. • Micro-macro aggregation equivalence is fully characterised and delimited. • Product t-norm uniquely preserves marginals for soft-label classification. • Confusion matrices span an orthogonal basis with m(m-1) degrees of freedom.
Keywords:
Classification performance measures
Multilabel classification
Evaluation framework
Micro-averaging
Macro-averaging
Cost-sensitive evaluation
Confusion matrix,

Journal

Information Fusion cover
Information Fusion
IF:
15.5
Papers:
4.2K
Citations:
2.7W

Organization

No organization information available