Skip to main content

Table 4 Classification performances of the selected classifiers.

From: A pilot study for development of a novel tool for clinical decision making to identify fallers among ophthalmic patients

Classifier

parameters

Accuracy

(95% CI)

Sensitivity

Specificity

AUC

AB

NI: 200; CF: 0.5; ML: 25

74.7 %

(71.1% - 77.4%)

75.8 %

73.7 %

82.3 %

C4.5

CF: 0.4; ML: 20

63.7 %

(60.0% - 66.9%)

57.9 %

69.5 %

65.5 %

RF

NT = 200; NV = 15

75.3 %

(71.7% - 77.9%)

72.6 %

77.9 %

86.2 %

  1. CI: Confidence Interval;
  2. AB: Adaboost;
  3. RF: Random Forest;
  4. NI: number of iteration;
  5. ML: minimum number of instances per leaf;
  6. CF: confidence factor for pruning;
  7. NT: number of trees;
  8. NF: number of randomly chosen features.