Using Machine Learning Techniques to Predict Learner Drop-out Rate in Higher Educational Institutions

Table 1

Summary of literature review.

Review	Mode	K-fold cross-validation	Feature selection method	Algorithms compared	Best classifier and metrics reported
[25]	Online MOOCs	Not stated	None	RF, GLM, GBM, MNET1, and MNET2	MNET1, accuracy = 91.57%
[26]	✓	10-Fold	✓	Deep learning, KNN, SVM, and DT	Deep learning, accuracy = 95.8%
[27]	✓	5-Fold	✓	DT	DT, f-measure = 76.3%
[28]	✓	Not stated	✓	GRU-RNN, XGBoost, GBDT, and RF	GRU-RNN, accuracy not stated
[29]	Traditional classroom	5-Fold	✓	RF, NN, SVM, and LOGIC	RF, accuracy = 94%
[30]	✓	Not stated	✓	RF and BDT	BDT, ROC value = 0.898
[31]	✓	10-Fold	✓	LR and DT	DT, accuracy not stated
[32]	✓	10-Fold	✓	DT, LR, RF, SVM, NB, and kNN	RF, f-measure = 0.975
[33]	✓	Not stated	✓	DT, LR, and NB	DT, ROC value = 0.94
[34]	✓	Not stated	✓	NB	NB, accuracy = 72%