Mortality prediction in ICU Using a Stacked Ensemble Model

Table 1

Example of related literature and technical comparison.

References	Technique	Characteristics
[1]	Weight decay RF, integrate the missing value analysis and likelihood ratio test.	Based on sparse data, robust calibration for large-scale dataset.
[7]	APACHE IV scoring system method	Weak calibration for new data records.
[10]	SOFA-based ML, (XGBoost, RF, SVM, and LR)	Comparison with four different ML approaches, incorporating the time information.
[12]	LR and ensemble techniques.	Poor performance for the nonlinear relationship
[13]	XGBoost	Improved accuracy and applicability; promising performance for nonlinear relationship; SHAP analysis for interpreting the model.
[16]	Deep learning(rnn)	Learn complex interactions from the data; reduced explainability for the model.