Table 6.
P-values of the Mann Whitney statistic to test the significant difference between the performance of RF with that of Bagging, Boosting, Logistic regression, kNN and Naïve Bayes classifiers in all the three encoding procedures under both balanced and imbalanced situations
| $D | EP | CLs | TPR | TNR | F (α = 1) | F (β = 2) | G-mean | WA | MCC |
|---|---|---|---|---|---|---|---|---|---|
| Balanced | P-1 | RF-BG | 0.343066 | 0.676435 | 0.272856 | 0.212122 | 0.185711 | 0.240436 | 0.272856 |
| RF-BS | 0.820063 | 0.939006 | 0.314999 | 0.795936 | 0.314999 | 0.383598 | 0.314999 | ||
| RF-LG | 0.001672 | 0.053092 | 0.002879 | 0.000725 | 0.005196 | 0.009082 | 0.005196 | ||
| RF-NB | 0.000242 | 0.002796 | 1.08E-05 | 1.08E-05 | 1.08E-05 | 0.000181 | 1.08E-05 | ||
| RF-KN | 0.053182 | 0.087051 | 0.028806 | 0.063013 | 0.035463 | 0.025581 | 0.028806 | ||
| P-2 | RF-BG | 0.41319 | 0.594314 | 0.356232 | 0.356232 | 0.277512 | 0.315378 | 0.356232 | |
| RF-BS | 0.837765 | 0.367844 | 0.968239 | 0.968239 | 0.842105 | 0.743537 | 0.842105 | ||
| RF-LG | 0.000275 | 0.000439 | 2.17E-05 | 2.17E-05 | 2.17E-05 | 2.17E-05 | 2.17E-05 | ||
| RF-NB | 0.000275 | 0.004216 | 2.17E-05 | 2.17E-05 | 2.17E-05 | 2.17E-05 | 2.17E-05 | ||
| RF-KN | 0.000376 | 0.000273 | 2.17E-05 | 2.17E-05 | 2.17E-05 | 0.000278 | 2.17E-05 | ||
| P-3 | RF-BG | 0.171672 | 0.879378 | 0.14314 | 0.14314 | 0.165494 | 0.15062 | 0.14314 | |
| RF-BS | 0.494174 | 0.381613 | 0.970512 | 0.528849 | 0.853428 | 0.820197 | 0.911797 | ||
| RF-LG | 0.000181 | 0.000181 | 1.08E-05 | 1.08E-05 | 1.08E-05 | 1.08E-05 | 1.08E-05 | ||
| RF-NB | 0.000182 | 0.000279 | 1.08E-05 | 1.08E-05 | 1.08E-05 | 0.000182 | 1.08E-05 | ||
| RF-KN | 0.000182 | 0.000181 | 1.08E-05 | 1.08E-05 | 1.08E-05 | 0.000182 | 1.08E-05 | ||
| Imbalanced | P-1 | RF-BG | 0.000269 | 0.000251 | 2.17E-05 | 2.17E-05 | 2.17E-05 | 0.000278 | 2.17E-05 |
| RF-BS | 0.000176 | 0.002555 | 0.000181 | 0.000181 | 0.000181 | 0.000178 | 0.000181 | ||
| RF-LG | 0.000263 | 0.000268 | 2.17E-05 | 2.17E-05 | 2.17E-05 | 0.000263 | 2.17E-05 | ||
| RF-NB | 0.000271 | 0.177338 | 2.17E-05 | 2.17E-05 | 2.17E-05 | 2.17E-05 | 2.17E-05 | ||
| RF-KN | 0.000175 | 0.025526 | 1.08E-05 | 1.08E-05 | 1.08E-05 | 0.000182 | 0.000179 | ||
| P-2 | RF-BG | 0.000179 | 0.000173 | 0.000182 | 0.000182 | 0.000182 | 0.000181 | 0.000182 | |
| RF-BS | 0.000181 | 0.000158 | 1.08E-05 | 1.08E-05 | 1.08E-05 | 0.000181 | 1.08E-05 | ||
| RF-LG | 0.00018 | 0.000178 | 1.08E-05 | 1.08E-05 | 1.08E-05 | 0.00018 | 1.08E-05 | ||
| RF-NB | 0.000182 | 0.733634 | 1.08E-05 | 1.08E-05 | 1.08E-05 | 0.000182 | 1.08E-05 | ||
| RF-KN | 0.000181 | 0.000174 | 1.08E-05 | 1.08E-05 | 1.08E-05 | 0.000182 | 1.08E-05 | ||
| P-3 | RF-BG | 0.000176 | 0.000168 | 0.000182 | 0.000182 | 0.000182 | 0.000181 | 0.000182 | |
| RF-BS | 0.000179 | 0.000149 | 1.08E-05 | 1.08E-05 | 1.08E-05 | 1.08E-05 | 1.08E-05 | ||
| RF-LG | 0.000179 | 0.000177 | 1.08E-05 | 1.08E-05 | 1.08E-05 | 0.000182 | 1.08E-05 | ||
| RF-NB | 0.000177 | 0.009082 | 1.08E-05 | 1.08E-05 | 1.08E-05 | 1.08E-05 | 1.08E-05 | ||
| RF-KN | 0.00018 | 0.00018 | 1.08E-05 | 1.08E-05 | 1.08E-05 | 0.000178 | 1.08E-05 |
$D data type, RF random forest, CLs classifiers, BG bagging, BS boosting, LG logistic regression, NB naïve bayes, KN K nearest neighbor