Table 5.

Performance statistics of the three classification models as well as diagnosis codes within each of the four health systems.

	Regularized Logistic Regression	Support Vector Machine	Random Forest	Coded Diagnosis
Health System 1
Accuracy	0.967	0.967	0.967	0.942
Recall	0.962	0.942	1.000	0.962
Precision	0.962	0.980	0.929	0.909
F1 score	0.962	0.961	0.963	0.935
Health System 2
Accuracy	0.967	0.967	0.967	0.940
Recall	0.985	0.985	0.985	0.985
Precision	0.941	0.941	0.941	0.889
F1 score	0.962	0.962	0.962	0.934
Health System 3
Accuracy	0.936	0.960	0.928	0.928
Recall	0.910	0.955	0.925	0.940
Precision	0.968	0.970	0.939	0.926
F1 score	0.938	0.962	0.932	0.933
Health System 4
Accuracy	0.923	0.933	0.904	0.913
Recall	0.951	1.000	1.000	0.951
Precision	0.867	0.854	0.804	0.848
F1 score	0.907	0.921	0.891	0.897