Skip to main content

Table 15 Comparison of the performance of Random Forest (RF) classifier having 10, 50 and 100 trees with different feature set considered at each split (1, 2, 4, 8, and 12) without using sampling

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

  No. of tree =10 No. of tree =50 No. of tree =100
F = 1 F = 2 F = 4 F = 8 F = 12 F = 1 F = 2 F = 4 F = 8 F = 12 F = 1 F = 2 F = 4 F = 8 F = 12
Sensitivity 45.82% 47.20% 48.35% 46.64% 45.44% 56.62% 56.34% 57.33% 55.84% 54.51% 58.39% 59.87% 59.09% 56.56% 54.41%
Specificity 90.03% 90.23% 90.58% 90.83% 90.85% 89.60% 89.81% 90.29% 90.48% 90.57% 89.50% 89.81% 90.21% 90.45% 90.48%
Precision 18.90% 20.77% 24.19% 26.91% 27.30% 13.61% 15.74% 20.41% 22.42% 23.42% 12.49% 15.53% 19.52% 22.08% 22.56%
F-score 26.76% 28.84% 32.24% 34.13% 34.11% 21.95% 24.61% 30.10% 31.99% 32.76% 20.58% 24.66% 29.35% 31.76% 31.90%
RMSE 0.3 0.3 0.3 0.3 0.31 0.29 0.29 0.29 0.29 0.30 0.29 0.29 0.29 0.29 0.29
AUC 0.76 0.77 0.77 0.77 0.76 0.81 0.81 0.81 0.81 0.80 0.81 0.81 0.82 0.81 0.81