Catching the Phish: Detecting Phishing Attacks Using Recurrent Neural Networks (RNNs)
Metadane
- Autorzy: L. Halgaš, I. Agrafiotis, J.R.C. Nurse
- Rok: 2020
- Źródło: Lecture Notes in Computer Science, Volume 11897, Springer, Pages 219–233
- DOI/Link: 10.1007/978-3-030-39303-8_17
- Status: to-read
- Pochodzenie: Wyekstrahowane z al-subaiey-web-ai-phishing-2024 ([26] - RNN classifier, Table 3 comparison)
- Tagi: to-read reference rnn recurrent-neural-network phishing-detection text-tokenization deep-learning
Notatki
Publikacja dodana automatycznie z bibliografii.
Kontekst cytowania w Al-Subaiey 2024:
- Referenced jako [26] w paper (Table 1, Table 3 comparisons)
- Method: Text tokenization → Recurrent Neural Network classifier
- Dataset: SpamAssassin + Enron + Nazario (multiple subsets)
- SA-JN: 4,572 spam, 6,951 ham
- En-JN: 9,962 spam, 10,000 ham
- Results: Accuracy 98.91%, Precision 98.74%, Recall 98.53%, F1-score 98.63%
- Comparison: Al-Subaiey 99.1% acc outperforms with simpler SVM approach
Technical Details:
- Preprocessing: tokenization (no mention of stop word removal in Al-Subaiey summary)
- RNN architecture (LSTM likely, though not specified in citation)
- Combined datasets strategy similar to Al-Subaiey’s approach
Significance:
- Demonstrates RNN effectiveness for sequential text processing
- Multiple dataset evaluation shows generalization
- Al-Subaiey’s finding: traditional ML competitive with RNNs for email classification
Dodaj PDF aby wygenerować pełne podsumowanie używając /summarize-paper halgas-rnn-phishing-detection-2020