Skip to main content
Advertisement
Browse Subject Areas
?

Click through the PLOS taxonomy to find articles in your field.

For more information about PLOS Subject Areas, click here.

< Back to Article

Fig 1.

Hybrid deep-learning architecture for cyberbullying detection.

This architecture shows the multimodal architecture combining CNN, BiLSTM, and Transformer encoders. The semantic consistency layer enforces alignment among modalities using the loss defined in Equation 2, improving contextual reliability before classification.

More »

Fig 1 Expand

Table 1.

Feature extraction components.

More »

Table 1 Expand

Table 2.

Experimental configuration and training parameters.

More »

Table 2 Expand

Fig 2.

Sequential workflow of the proposed framework.

More »

Fig 2 Expand

Fig 3.

Training and validation curves.

More »

Fig 3 Expand

Fig 4.

Confusion matrices for CAVD and SocialVidMix.

More »

Fig 4 Expand

Fig 5.

ROC (AUC) curves for proposed vs baseline models.

More »

Fig 5 Expand

Fig 6.

Precision–recall (PR) curves.

More »

Fig 6 Expand

Table 3.

Dataset characteristics and annotation reliability.

More »

Table 3 Expand

Table 4.

Performance comparison on CAVD and SocialVidMix datasets.

More »

Table 4 Expand

Table 5.

Cross-platform generalization results.

More »

Table 5 Expand

Table 6.

Ablation study results on CAVD dataset.

More »

Table 6 Expand