Random forest algorithms yielded the best classification performance in K–8 MMLA studies comparing multiple machine learning models¶

Claim · All claims
Evidence · 1 study · review r2 · q2
Currency · newest study 2025 · checked against its sources 2026-10-20 (automated, 1 of 1 entry) · not reviewed by a person · recheck by 2028: tool and model accuracy changes quickly

Subclaims¶

q2 i? In studies comparing classifiers, random forest was chosen for best performance, including detecting guessing behaviours with motion features. → Caskurlu 2025

Evidence¶

Caskurlu 2025¶

Caskurlu, S., Ocak, C., & Dai, C.-P. (2025). The Scope of Multimodal Learning Analytics in K–8: A Systematic Review. Journal of Learning Analytics, 12(2), 224–236. https://doi.org/10.18608/jla.2025.8505

q2 · i? · review · r2

The review reports that Shin et al. (2020) compared logistic regression, support vector machine, decision tree, and random forests with baseline classifiers using synthetic minority over-sampling. Similarly, in Moon et al. (2022), after testing kNN, random forest, decision tree, and logistic regression, "random forest was chosen as a classification model because it resulted in the best performance."

"Their results showed th at random for est yielded the best performance in detecting guessing behavio urs with motion features."

Discussion¶

Related Claims¶

• A support vector machine classifier outperformed decision tree and random forest models in predicting cognitive engagement levels of online discussion posts — reports the opposite
• A random forest classified flagged versus non-flagged word problems with the best accuracy of five tested models (AUC = 0.75) — a narrower finding that bears on this claim
• Random Forest is the most commonly used NLP-supporting algorithm in LA, while deep learning models (LSTM, BERT) have been widely adopted since 2021 — related
• CatBoost achieved the best classification performance among tested algorithms for predicting course completion risk (F-measure .77, accuracy 78%, AUC .87) — reports the opposite
• All three trained classifiers outperformed the zero-rule baseline (28.4% accuracy) for classifying cognitive engagement in discussion posts — related
• SVM and random forest classifiers showed consistent performance in classifying ethnic bias, with F1-scores of 0.71 and 0.70 on the test set — related
• Simple classifiers (Logistic Regression, Naive Bayes) outperformed tree-based models on this small dataset — reports the opposite
• On the TU/e CS overlap dataset, supervised NLP classifiers (Random Forest, XGBoost) outperform thresholding and zero-shot LLM baselines at detecting course overlap — a narrower finding that bears on this claim
• A Random Forest classifier reached 89.09% test accuracy with 0.89 macro F1 across three stress-severity levels, exceeding an SVM comparison (88.48%) and a reported Random Forest benchmark (87.93%) on similar data — a narrower finding that bears on this claim
• On the CASED challenge test set, all participating engagement-prediction methods converge near random-chance performance, with the best F1 Macro at 0.52 — related
• In a student project on laser-drilled hole diameter detection, GPR significantly outperformed MLP in detection accuracy and training speed — related
• Decision trees achieve competitive accuracy with black-box models (LR, RF, SVM), which performed slightly better but lack interpretability — reports the opposite
• Decision tree classifiers trained on slide interaction features predict end-of-module outcomes with high accuracy across four K-12 programming courses — related
• Machine learning methods outperformed logistic regression in classification accuracy across reviewed studies — a broader claim this one bears on
• A stacking ensemble of LG, SVM, NB, KNN and XGB achieved the best fine-tuning F1-score (0.88), outperforming individual models such as SVM (0.86) — related
• Predictive models trained on one semester's offering of a course identified at-risk students in the subsequent semester with high prediction accuracy — related
• An LSTM-based strategy temporal analytics framework outperforms random forest, SVM, and majority-class baselines for stealth assessment predictive accuracy — reports the opposite
• Students not on track for college enrollment and persistence can be classified with about 90 percent accuracy using a small set of predictors — related
• Random Forest reaches CV F1 = 0.924 on the psychographic survey dataset, but the article presents this only as a localized exploratory pilot — a narrower finding that bears on this claim
• Random-forest models trained on a prior semester showed significantly lower AUC when tested on a new semester's data without retraining — related