Neural Network-Based Coronary Heart Disease Risk Prediction Using Feature Correlation Analysis

Table 3

Results of feature sensitivity analysis.

Features	Sensitivity	NNoutput-NNoutput(xi + δ)	Rank
NNk(X)	0.815
NNk(X(xi−age))	0.734	0.081	2
NNk(X(xi−sex))	0.726	0.008	11
NNk(X(xi−BMI))	0.769	0.038	5
NNk(X(xi−To_chole))	0.677	0.100	1
NNk(X(xi−HDL))	0.703	0.013	8
NNk(X(xi−SBP))	0.729	0.073	3
NNk(X(xi−DBP))	0.693	0.049	4
NNk(X(xi−triglyceride))	0.753	0.013	7
NNk(X(xi−hemoglobin))	0.796	0.006	12
NNk(X(xi−TD))	0.806	0.003	13
NNk(X(xi−CRF))	0.802	0.010	10
NNk(X(xi−H_B))	0.813	0.001	15
NNk(X(xi−H_C))	0.813	0.001	16
NNk(X(xi−cirrhosis))	0.812	0.002	14
NNk(X(xi−smoking))	0.802	0.012	9
NNk(X(xi−diabetes))	0.786	0.024	6