Classical machine learning models outperform Hungarian transformers slightly in overall performance (71% vs 68% average score) for classifying reflection levels in student essays, though transformers handle rare classes better.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Automatic Reflection Level Classification in Hungarian Student Essays
Classical machine learning models outperform Hungarian transformers slightly in overall performance (71% vs 68% average score) for classifying reflection levels in student essays, though transformers handle rare classes better.