Table 9.

EDR selected studies.

Authors Name and Year	Methods	Results	Authors Suggestions/Conclusions
Cui et al., (2021) [120]	Extreme Gradient Boost (XGBoost) algorithm	Accuracy = 96.2, Precision = 86.5, Recall = 83.0	ML methods showed promise for forecasting multiclass issues, such as varying therapies depending on EDRs.
Kang et al., (2022) [121]	RF, ANN, CNN, GBDT, SVM, LR, LSTM	Accuracy = 92%, F1-score = 90%, precision = 94%, recall = 87%	ML is strongly recommended as a decision-making aid for dental practitioners in the early diagnosis and treatment of tooth caries
Chen, (2021) [122]	NLP	F1-score 83% and 88%	The NLP workflow might be used as the initial stage to training data-based models with structured data.