ClickGuard adaptively fuses BERT semantic embeddings and structural features with a hybrid CNN-BiLSTM model to reach 96.93% accuracy on clickbait detection while using LIME and permutation feature importance for interpretability.
Explainable hybrid word representations for sentiment analysis of financial news,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.CL 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
ClickGuard: A Trustworthy Adaptive Fusion Framework for Clickbait Detection
ClickGuard adaptively fuses BERT semantic embeddings and structural features with a hybrid CNN-BiLSTM model to reach 96.93% accuracy on clickbait detection while using LIME and permutation feature importance for interpretability.