Imbalanced class distribution and performance evaluation metrics: A systematic review of prediction accuracy for determining model performance in healthcare systems

Michael Owusu-Adjei, James Ben Hayfron-Acquah, Twum Frimpong, Gaddafi Abdul-Salaam, Henry Horng-Shing Lu, Anat Reiner-Benaim, Henry Horng-Shing Lu, Anat Reiner-Benaim, Henry Horng-Shing Lu, Anat Reiner-Benaim · 2023

Read Imbalanced class distribution and performance evaluation metrics: A systematic review of prediction accuracy for determining model performance in healthcare systems by Michael Owusu-Adjei, James Ben Hayfron-Acquah, Twum Frimpong, Gaddafi Abdul-Salaam, Henry Horng-Shing Lu, Anat Reiner-Benaim, Henry Horng-Shing Lu, Anat Reiner-Benaim, Henry Horng-Shing Lu, Anat Reiner-Benaim (2023) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-10688675.txt.

Description

Michael Owusu-Adjei and co-authors review how skewed class counts distort accuracy scores for health prediction models, in work from twenty twenty-three. Most studies crown the best model by a plain accuracy score. That habit ignores how uneven the outcome classes are in the data. In health data, rare outcomes often matter most and get swamped. A model can score high while missing nearly every rare case.

The review asks how class skew bends accuracy versus balanced accuracy. Balanced accuracy weights each class fairly, no matter its size. Past works that use skewed data with both scores are gathered and weighed. The aim is to place the debate in real health and field settings. Findings should guide which score to trust when classes run uneven. Better scoring can stop weak models from looking strong on paper. Care systems that act on such models gain from the fix.

Catalog record

CREATORMichael Owusu-Adjei, James Ben Hayfron-Acquah, Twum Frimpong, Gaddafi Abdul-Salaam, Henry Horng-Shing Lu, Anat Reiner-Benaim, Henry Horng-Shing Lu, Anat Reiner-Benaim, Henry Horng-Shing Lu, Anat Reiner-Benaim
YEAR2023
COLLECTIONPAPERS
FORMATTXT
FILE SIZE67 KB
VERSION1.0
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC10688675

Browse the full library · About the Tower of Babel