Research Article

Machine Learning Approach to Extract Diagnostic and Prognostic Thresholds: Application in Prognosis of Cardiovascular Mortality

Table 3

Performance of classifiers.

ClassifiersSensitivitySpecificityAccuracy

I f systolic ARV ≥ 9.6
then 1 Else 0
55.7%60.4%59.9%
If systolic WBP ≥ 134.6
then 1 Else 0
52.5%58.8%58.08%
If systolic ARV ≥ 9.6
and systolic WBP ≥ 137
then 1 Else 0
36.1%80.0%75.1%
If systolic ARV ≥ 9.6
and systolic WBP ≥ 138.6
and cholesterol ≥ 5.5
then 1 Else 0
8.2%93.3%83.8%
If systolic ARV ≥ 10.4
and systolic WBP ≥ 139.8
and BMI ≥ 27.3
then 1 Else 0
9.8%93.3%84.0%
If systolic ARV ≥ 9.6
and systolic WBP ≥ 137
and diastolic WBP ≥ 78.4
then 1 Else 0
22.9%87.5%80.4%
Naïve Bayes11.48%95.92%86.57%