Synapse
⌘+K
Synapse
PulseExploreClubsResearchersJournals
Instagram
HomeClubsExplore
March 28, 2026Korean Journal of RadiologyOpen Access

Key Measures for Evaluating Diagnostic Accuracy in Multi-Class Classification: An Overview and Simulation-Based Comparison

View Full Paper
Ask AI
Bookmark
Share

Authors

LRLeeha RyuYonsei UniversityKHKiwan HanUniversity Medical Center UtrechtIJInkyung JungKorea Advanced Institute of Science and Technology

Discussion

Loading...

Member takes

Overview

Simulation study evaluates diagnostic accuracy measures in multi-class systems, suggesting best practices for application.

Key Points

  • This study aims to evaluate and compare diagnostic accuracy metrics for multi-class classification using comprehensive simulations.
  • Overview of commonly used accuracy metrics for multi-class classification.
  • Conducted a simulation study across balanced and imbalanced sample sizes.
  • Evaluated several scenarios, including three and five-class settings.
  • Assessed performance of each metric based on bias and confidence interval coverage.
  • Most metrics showed stable and unbiased performance under balanced conditions.
  • Greater bias observed under imbalanced conditions, particularly for micro-averaged receiver operating characteristic curve area.
  • The M-index and polytomous discrimination index provided more stable performance across scenarios.

Cite This Study

Ryu et al. (2026) studied this question.

synapsesocial.com/papers/69c76fff8bbfbc51511e059ahttps://doi.org/10.3348/kjr.2025.1447
View Full Paper
Ask AI
Bookmark
Share

Also Consider

Synapse has enriched 5 closely related papers on similar clinical questions. Consider them for comparative context:

  1. 1Polytomous diagnosis of ovarian tumors as benign, borderline, primary invasive or metastatic: development and validation of standard and kernel-based risk prediction models2010 · 38 citations
  2. 2Domain-adapted Large Language Models for Classifying Nuclear Medicine Reports2023 · 28 citations
  3. 3Real-time, automatic, open-source sleep stage classification system using single EEG for mice2021 · 26 citations
  4. 4LI-RADS: Current Status and Future Directions2024 · 18 citations
  5. 5Development, deployment, and feature interpretability of a three-class prediction model for pulmonary diseases2025 · 3 citations