Table 5.

Regression method’s Spearman coefficient as a function of feature set. The Spearman coefficient is calculated for each regression technique with and without its respective feature selection method. The EMR data that is used is constrained to up to 48 hours before the biomarker measurement and 1 hour after.

Regression Model	All features (μ,σ)	EMR (μ,σ)	Biomarkers (μ,σ)
Logistic Regression	0.62, 0.10	0.38,0.13	0.52, 0.11
Logistic Regression w/ feature selection	0.60,0.10	0.36, 0.14	0.52, 0.11
SVM	0.63, 0.09	0.45, 0.12	0.53, 0.11
SVM w/ feature selection	0.59. 11	0.47, 0.11	0.46, 0.12
Random Forest	0.69, 0.08	0.50, 0.11	0.56, 0.10
Random Forest w/ feature selection	0.69,0.07	0.51, 0.11	0.57,0.10
Adaboost	0.67, 0.08	0.47, 0.12	0.57, 0.10
Adaboost w/ feature selection	0.62, 0.09	0.47, 0.12	0.56, 0.10