Skip to main content

Table 10 Comparison of the performance of Bayesian Network classifier (BN) using different search algorithms: K2, Hill Climbing, Repeated Hill Climber, LAGD Hill Climbing, TAN, Tabu and Simulated Annealing using Sampling

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

  K2 Hill Climbing Repeated Hill Climber LAGD Hill Climbing TAN Tabu Simulated Annealing
Sensitivity 37.44% 37.44% 37.44% 47.65% 60.07% 37.59% 55.20%
Specificity 93.31% 93.31% 93.31% 91.55% 91.02% 93.20% 91.23%
Precision 52.11% 52.11% 52.11% 33.76% 27.32% 51.10% 29.71%
F-score 43.57% 43.57% 43.57% 39.52% 37.56% 43.31% 38.63%
RMSE 0.34 0.34 0.34 0.34 0.28 0.34 0.29
AUC 0.82 0.82 0.82 0.81 0.84 0.81 0.84