Detection of risk levels in optical mark recognition exams using categorical feature-based machine learning models

Yasin Sancar · 2025

Read Detection of risk levels in optical mark recognition exams using categorical feature-based machine learning models by Yasin Sancar (2025) 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-12739130.txt.

Description

Yasin Sancar builds machine learning models that flag risk levels from optical mark exam sheets. Optical systems score large exams fast, but they ignore how answers were marked. Blank items, double marks, and repeated patterns can hint at conduct or care issues.

Sancar used expert-set risk labels with five levels as the target. All inputs were category features from the answer sheets. Models were tested with stratified cross-validation on held-out data. CatBoost scored best, well above a rule-based baseline built from expert cutoffs. A merged three-level version scored higher still. Confusion was mostly between nearby risk levels. The study shows marked sheets hold more signal than scores alone.

Catalog record

CREATORYasin Sancar
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE53 KB
VERSION1.0
LICENSECC BY 4.0
RIGHTSCC BY
SOURCEEurope PMC

Source: https://europepmc.org/articles/PMC12739130

Browse the full library · About the Tower of Babel