Skip to main content

Table 3 Performance of meta classification techniques compared to bagged J48.

From: Enhancing navigation in biomedical databases by community voting and database-driven text classification

 

Bagging

AdaBoost

MultiBoost

Dataset

J48

PART

J48

PART

J48

PART

CA

92.1 (5.0)

91.9 (5.5)

93.1 (5.4)

92.0 (5.9)

92.5 (4.8)

91.3 (6.0)

CVD

91.8 (4.6)

91.1 (5.6)

91.9 (5.0)

90.5 (5.8)

91.9 (5.0)

91.6 (5.4)

DM

96.4 (3.6)

96.2 (3.7)

97.3 (3.6)

97.0 (3.6)

97.0 (3.3)

96.9 (3.5)

APO

98.3 (3.1)

98.3 (3.1)

98.9 (2.5)

98.3 (2.8)

98.6 (2.7)

97.9 (3.4)

ANG

95.3 (4.2)

95.6 (4.2)

96.0 (3.6)

94.9 (4.1)

95.1 (4.2)

95.1 (4.3)

MI

94.5 (3.9)

94.6 (3.5)

95.4 (3.8)

95.1 (3.9)

94.9 (4.1)

94.4 (3.8)

BD

90.7 (5.3)

89.6 (6.3)

91.4 (4.3)

90.1 (6.1)

91.2 (5.1)

89.6 (6.1)

  1. F-Measure (in %) and standard deviation. No statistically significant degradation or improvement was observed.