“…The second most popular algorithm was NN (14, 33%): many studies used multiple hidden layers based deep learning techniques [ 60 , 69 – 71 , 77 , 79 , 80 , 85 – 87 ] (e.g., recurrent NN, convolutional NN, deep NN, and ensemble of DL networks), while a few other studies either used one hidden layer [ 58 , 60 , 68 ] or did not specify the number of layers [ 49 , 66 ]. Regularized logistic regression (12, 28%), including Least Absolute Shrinkage and Selection Operator (LASSO) regression [ 53 , 64 , 65 , 67 , 70 , 71 , 78 – 80 ] (L1 regularization), ridge regression [ 64 , 70 , 71 , 80 ] (L2 regularization) and elastic-net [ 49 , 72 , 81 ]were third most used ML algorithm, followed by Support Vector Machine (SVM) [ 54 , 60 , 63 , 65 , 66 , 70 , 71 , 82 – 84 ] (10, 23%). The other less commonly used ML algorithms included naïve Bayes network [ 49 , 54 , 70 , 84 ], K-Nearest Neighbors (KNN) algorithm [ 54 , 65 ], ensemble of methods [ 50 , 67 , 84 ], ,and Bayesian Model averaging [ 49 ].…”