Contents
v
MIPSVM Classifier with Existing Classifiers
135
IV
Appendices
149
A Software Tools
151
A.1 Zereal
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 151
A.2 Incridge and pIncridge
. . . . . . . . . . . . . . . . . . . . . . . . . 152
A.3 Other software developed
. . . . . . . . . . . . . . . . . . . . . . . 152
A.3.1
Browsim
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
A.3.2
jfipa
. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 152
B Statistical Analysis
153
B.1 Factorial Design for Zereal Performance
. . . . . . . . . . . . . . . 153
B.1.1
Factorial Design Experimental Data
. . . . . . . . . . . . . 153
B.1.2
Analysis of Variance (ANOVA)
. . . . . . . . . . . . . . . . 154
B.2 Classification Accuracy - UCI Datasets
. . . . . . . . . . . . . . . . 156
B.3 Computational Performance - UCI Datasets
. . . . . . . . . . . . . 156
B.4 Data for Computational Performance - Zereal and Web data
. . . 156
B.5 Paired T-Tests for Classification Accuracy
. . . . . . . . . . . . . . 157
B.5.1
Classification Data C4.5 and Naive Bayes
. . . . . . . . . . 157
B.5.2
MIPSVM and C4.5 Comparison
. . . . . . . . . . . . . . . 157
B.5.3
MIPSVM and Naive Bayes Comparison
. . . . . . . . . . . 158
B.5.4
Classification Data Log. Regr. and Vot. Perc.
. . . . . . . . 158
B.5.5
MIPSVM and Logistic Regression Comparison
. . . . . . . 159
B.5.6
MIPSVM and Voted Perceptron Comparison
. . . . . . . . 159
Bibliography
161
Index
172
<<
-
<
-
>
-
>>