Skip to main content

Table 9 Comparison of the performance of Naïve Bayesian classifier (BC) using three different Weka options for handling continuous attributes: single normal, kernel estimation and supervised discretization without using Sampling

From: Comparison of machine learning techniques to predict all-cause mortality using fitness data: the Henry ford exercIse testing (FIT) project

  Single Normal kernel Estimation Supervised Discretization
Sensitivity 35.73% 41.25% 37.71%
Specificity 93.22% 92.17% 93.23%
Precision 51.89% 40.79% 51.32%
F-score 42.32% 41.02% 43.47%
RMSE 0.35 0.32 0.34
AUC 0.81 0.81 0.82