158
Statistical Analysis
Paired T for MIPSVM - C4.5
N
Mean
StDev
SE Mean
MIPSVM
18
58.64
28.60
6.74
C4.5
18
63.52
27.39
6.46
Difference
18
-4.87
11.40
2.69
95% CI for mean difference: (-10.54; 0.80)
T-Test of mean difference = 0 (vs not = 0):
T-Value = -1.81
P-Value = 0.088
B.5.3
MIPSVM and Naive Bayes Comparison
The P-Value based on the T-test is 0.013, this is significant according to a 5%
significance level, hence the experiments showed that MIPSVM is more accurate
than Naive Bayes on the experiment data sets used.
Paired T-Test and CI: MIPSVM; Naive Bayes
Paired T for MIPSVM - Naive Bayes
N
Mean
StDev
SE Mean
MIPSVM
18
58.64
28.60
6.74
Naive Bayes
18
49.61
33.26
7.84
Difference
18
9.04
13.76
3.24
95% CI for mean difference: (2.20; 15.88)
T-Test of mean difference = 0 (vs not = 0):
T-Value = 2.79
P-Value = 0.013
B.5.4
Classification Data Log. Regr. and Vot. Perc.
Dataset
MIPSVM
LogReg. Vot. Perc.
P0
46.5116 52.8241 45.8333
P1
84.4444 83.6957 83.1522