Table 2.
Result table for Set 1 with four classifier algorithms Naïve Bayes, Random Forest, J48 and SMO
|
Classifiers |
TP% |
FP% |
TN% |
FN% |
Accuracy% |
Average |
Average |
Average |
|---|---|---|---|---|---|---|---|---|
| Precision% | Recall% | ROC | ||||||
|
Naïve Bayes |
69.9 |
42.1 |
57.9 |
30.1 |
64.59 |
67.8 |
69.89 |
71.90% |
|
Random Forest |
83.7 |
21.7 |
78.3 |
16.3 |
79.81 |
79.5 |
78.3 |
89.2% |
|
J48 |
79 |
27 |
73 |
21 |
76.35 |
78.8 |
78.96 |
77.20% |
| SMO | 74 | 34.6 | 65.4 | 26 | 70.24 | 73.1 | 74.02 | 77.10% |
Random Forest showed the Best accuracy with 79.81% and ROC Area of 89.2%. It also has high True Positive (TP) rate with low False Positive (FP) rate.