Table 3.

Best hyper-parameters for ML algorithm modeling in prediction of readmission.

Num	Algorithms	Hyper-parameters	f-score
1	HistGradientBoostingClassifier	‘verbose’ = 2, ‘random_state’ = 999, ‘max_leaf_nodes’ = 62, ‘max_iter’ = 150, ‘max_depht’ = 7, ‘learning rate’ = 0.1	93.7
2	BaggingClassifier	‘verbose’ = 2, ‘random_state’ = 999, ‘n_estimation’ = 12, ‘max-samples’ = 0.5, ‘bootstrap’ = ‘true’	91.28
3	MLP Classifier	‘Learning rate’ = ‘constant’, hidden_layer_size’ = (100,100,100), ‘alpha’ = 0.05, ‘activation’ = ‘rulo’	91.07
4	SVM (kernel = linear)	C = 100,G = 0.0001	90.09
5	SVM (kernel = RBF)	C = 10, G = 0.001	89.24
6	XG Boost Classifier	‘min_chid_weigh’ = 1′max_depht’ = 12,‘learning_rate’ = 0.1, ‘gamma’ = 0.4, ‘colsample_bytree’ = 0.3	89.01
7	K Nearest Neighbor Classifier	K = 3, ‘n_jobs’ = −1, ‘algorithm’ = ‘auto’	87.00