Interpretable Machine Learning to Optimize Early In-Hospital Mortality Prediction for Elderly Patients with Sepsis: A Discovery Study

Table 1

Hyperparameters and tuning strategies for each model in the study.

Model	Hyperparameters	Grid-search setting
K-nearest neighbors	Number of neighbors	5, 10
	Neighbor weight	Uniform, distance
Decision tree	Maximum depth	10, 50, 100
Random forest	Number of estimators	10, 50, 100
	Maximum depth	10, 50, 100
Logistic regression	Cs	0.1, 1, 10, 100, 1000
LGBM	Learning rate	0.01, 0.1, 1
	Number leaves	5, 10, 15, 20
	Maximum depth	3, 5, 7, 9, 11
XGBoost	Learning rate	0.01, 0.1, 1
	Number leaves	5, 10, 15, 20
	Maximum depth	3, 5, 7, 9, 11