Research Article
Nature-Inspired-Based Approach for Automated Cyberbullying Classification on Multimedia Social Networking
Table 2
Results of predicting the CB with 60% training data with information gain.
| Statistical parameters | NB | LR | RF | SVM | ANN | ANN-DRL |
| Accuracy | 56.890 | 57.190 | 59.280 | 59.541 | 60.901 | 81.688 | F-measure | 39.614 | 41.714 | 52.948 | 53.108 | 55.479 | 84.869 | G-mean | 73.755 | 73.986 | 75.486 | 75.526 | 75.936 | 86.790 | MAPE | 29.550 | 26.609 | 25.209 | 22.628 | 22.048 | 17.346 | Sensitivity | 62.962 | 66.473 | 74.376 | 86.760 | 87.420 | 97.464 | Specificity | 75.396 | 75.586 | 79.097 | 79.117 | 80.488 | 81.328 |
|
|