Automatic Reflection Level Classification in Hungarian Student Essays

2026-05-04

AI Summary

This study addresses the time-consuming and subjective nature of manual assessment of reflective essays written by Hungarian-speaking students by introducing the first large-scale, expert-annotated dataset of Hungarian reflective writing (comprising 1,954 essays) and advancing the first automated approach for reflection-level classification in this language. The authors systematically tackle class imbalance through techniques including class weighting, oversampling, data augmentation, and alternative loss functions, and evaluate both classical methods—such as TF-IDF and semantic embeddings—and a Hungarian-specific Transformer model. Experimental results show that classical approaches achieve a balanced accuracy of 71%, while the Transformer demonstrates superior generalization on minority reflection classes, attaining 68% accuracy. These findings highlight the complementary strengths of traditional and deep learning methods in low-resource, imbalanced text classification settings.

Abstract

Reflective thinking is a key competency in education, but assessing reflective writing remains a time-consuming and subjective task for education experts. While automated reflective analysis has been explored in several languages, Hungarian language was not researched extensively. In this paper, we present the first comprehensive study on automatic reflection level classification in Hungarian student essays. We used a large, expert-annotated Hungarian dataset consisting of 1,954 reflective essays collected over multiple academic years and labeled on a four-level reflection scale. We investigate two approaches: (1) classical machine learning models using TF-IDF and semantic embedding features, and (2) Hungarian-specific transformer models fine-tuned for document-level reflection classification. To address the strong class imbalance in the dataset, we systematically examine class weighting, oversampling, data augmentation, and alternative loss functions. An extensive ablation study is conducted to analyze the contribution of each modeling and balancing strategy. Our results show that shallow machine learning models with appropriate feature engineering achieve strong overall performance, reaching up to 71% overall score averaged over accuracy, F1-score, and ROC AUC metrics, while transformer-based models achieve slightly lower overall score (68%) averaged over the same metrics, but demonstrate better generalization on minority reflection classes. These findings highlight the continued relevance of classical methods for low-resource settings and the robustness of transformer models for imbalanced classification. The proposed dataset and experimental insights provide a solid foundation for future research on automated reflective analysis in Hungarian and other morphologically rich languages.

Problem

• reflective writing
• automatic classification
• Hungarian language
• class imbalance
• educational assessment

Innovation

• reflection level classification
• Hungarian language
• class imbalance
• transformer models
• machine learning