Nature-Inspired-Based Approach for Automated Cyberbullying Classification on Multimedia Social Networking

Table 5

Results of predicting the CB with 60% training data with χ².

Statistical parameters	NB	LR	RF	SVM	ANN	ANN-DRL
Accuracy	57.480	60.091	62.452	63.822	67.063	85.900
F-measure	67.963	68.013	68.904	70.024	75.056	80.618
G-mean	44.725	57.650	60.681	45.926	77.307	87.200
MAPE	20.597	17.926	17.836	12.984	11.654	10.504
Sensitivity	77.447	80.128	80.218	85.059	86.400	87.550
Specificity	74.606	77.487	78.427	81.588	83.589	85.680