Table 1.

Differences between our work and previous machine-learning-based studies.

Author(s)	Year	Method(s)	Differences with our work
Khazali et al.16	2019	Fuzzy DT	Relying on only one model, smaller dataset, not considering previous production mechanism, different ML models
Giro et al.17	2019	Naïve Bayes	Using only one model, small dataset, fewer EOR methods, not considering previous production mechanism, incomplete test subset, different ML models
Cheraghi et al.6	2021	ANN, DT, NB, RF	Compared four models, but did not integrate their outputs, smaller dataset, not considering previous production mechanism, different ML models
Su et al.18	2023	RF, ANN, NB, DT, SVM	No integration of the models, smaller dataset, fewer EOR methods, not considering previous production mechanism, different ML models
Tabatabaei et al.19	2023	ANN-PSO, ANN-SSA	Used only one model, limited number of samples, not considering previous production mechanism, different ML models
Chavan et al.10	2023	ANN, KNN, SVM, RF, NB	Compared five models, but did not integrate their outputs, smaller dataset, not considering previous production mechanism, different ML models