CPSC 633 - Machine Learning
Project #4 – Na´ve Bayes Classifier 

Due 11/27/00

 

Unmil P. Karadkar

 

 

Database

DecisionTree

Perceptron
eta=0.01

 

Neural Network
eta = 0.05

Na´ve Bayes Classifier

 

1 Layer, #attrib nodes

1 hidden node

 

Breast Cancer

 -

74.88 ▒ 10.584

 73.76 ▒ 8.00

 72.47 ▒ 10.52

73.4 ▒ 8.86

 

Census

 -

 -

 

 

80.5 ▒ 3.16

 

Credit

 -

86.18 ▒ 3.47

 86.51 ▒ 5.09

 86.05 ▒ 3.81

78.5 ▒ 6.15

 

Diabetes

69.27

76.10 ▒ 6.07

 76.77 ▒ 5.74

 76.33 ▒ 5.68

75.44 ▒ 6.33

 

Exactly2

74.28 ▒ 4.87

75.56 ▒ 5.2

 84.6  ▒ 7.34

 76.48 ▒ 4.17

75.8 ▒ 4.04

 

Exactly3

69.56 ▒ 7.45

68.92 ▒ 5.29

 99.67 ▒ 0.84

 69.04 ▒ 7.76

69.4 ▒ 4.51

 

Glass

83.00

73.72 ▒ 13.44

 78.30 ▒ 15.46

 73.85 ▒ 16.48

71.26 ▒ 10.78

 

Heart

75.67

82.57 ▒ 7.13

 82.83 ▒ 8.98

 83.58 ▒ 5.58

82.3 ▒ 9.0

 

Iris

97.36

92.94 ▒ 10.53

 93.07 ▒ 8.10

 85.08 ▒ 23.62

94.16 ▒ 5.96

 

m-of-n

98.8 ▒ 1.34

97.16 ▒ 3.76

 99.87 ▒ 0.43

 99.74 ▒ 0.77

90.36 ▒ 3.78

 

Promoters

 -

81.70 ▒ 21.73

 

 

80.34 ▒ 17.03

 

Rotate

97.60

88.06 ▒ 4.15

 

 

91.02 ▒ 2.96

 

Secondary

 -

 -

 

 

62.2 ▒ 1.35

 

Voting

96.29

95.03 ▒ 4.35

 95.53 ▒ 3.88

 94.98 ▒ 4.31

91.18 ▒ 5.98