Journal of Analytical and Applied Computer Sciences | Volume 2 Issue 2 | Pages: 32-37 | Doi : 10.67914/jaacs/rsa/2.2/32-37

Research Article

OPEN ACCESS | Published on : 02-Sep-2026

An Intelligent Intrusion Detection Model for Prevention of Social Engineering Attacks on Social Media Network Platforms Using CNN-RNN-LSTM

• Sulaiman Ahmad
• Department of Computer Science, Faculty of Science, Federal University Kashere, Gombe State, Nigeria.

Peter Buba Zirra
• Department of Computer Science, Faculty of Science, Federal University Kashere, Gombe State, Nigeria.

Abstract

The increasing use of social media network platforms has created an attractive environment for social engineering attacks, where attackers exploit human trust, fear, urgency and other psychological cues rather than relying exclusively on technical vulnerabilities. This study presents an Intelligent Intrusion Detection Model (IIDM) for detecting social engineering attacks in textual social media content using a hybrid Convolutional Neural Network–Recurrent Neural Network–Long Short-Term Memory (CNN-RNN-LSTM) architecture. Natural Language Processing (NLP) techniques were applied to prepare and represent social media text, while the hybrid architecture was designed to combine local linguistic feature extraction with sequential/contextual pattern modelling. A cross-platform dataset containing 20,000 benign and malicious social media posts from Facebook and Twitter was cleaned, normalized and tokenized, with Synthetic Minority Over-sampling Technique (SMOTE) used to address class imbalance. The dataset was divided into 80% training and 20% testing sets. The model was trained using categorical cross-entropy and the Adam optimizer, with batch normalization, dropout and early stopping used to improve generalization. Experimental results showed an overall accuracy of 95%, macro precision of 0.95, macro recall of 0.89, macro F1-score of 0.91 and ROC-AUC of 0.89. Class-wise analysis showed strong benign classification, with precision of 0.95, recall of 0.99 and F1-score of 0.97, while the malicious class achieved precision of 0.94, recall of 0.78 and F1-score of 0.85. The results indicate that the hybrid model provides a practical approach for detecting social engineering threats in unstructured social media text and improves upon the social-media baseline reported in the study.

Keywords

Social engineering, intrusion detection, social media, natural language processing, CNN; RNN-LSTM, deep learning, phishing

References

• Abuali, K. M., Nissirat, L., & Al-Samawi, A. (2023). Intrusion detection techniques in social media cloud: Review and future directions. Wireless Communications and Mobile Computing, 2023, Article 6687023.  https://doi.org/10.1155/2023/6687023

  Adekunle, T. S., Lawrence, M. O., Alabi, O. O., Ebong, G. N., Ajiboye, G. O., & Bamisaye, T. A. (2024). The use of AI to analyze social media attacks for predictive analytics. American Journal of Operations Management and Information Systems, 9(1), 17–24.

  Alenezi, M., & Eisa, M. A. (2022). Comparative evaluation of machine learning techniques for intrusion detection systems. Computers & Security, 112, 102510. https://doi.org/10.1016/j.cose.2022.102510.

  Othman, A. M. H., Ab Razak, M. F., Firdaus, A., Tahaei, H., & Gheisari, M. (2026). A hybrid CNN–LSTM model for IoT intrusion detection: A robustness analysis across datasets. Future Internet, 18(7), 345.

  Alsubhi, A., & Tawalbeh, L. (2025). CNN-LSTM with attention for IoT intrusion detection.

  Etuh, E., Bakpo, F. S., & Agozie, E. H. (2021). Social media network attacks and their preventive mechanisms: A review. Computer Science & Information Technology, 11(4), 59–72. https://doi.org/10.5121/csit.2021.112405.

  Altunay, H. C., & Albayrak, Z. (2023). A hybrid CNN + LSTM-based intrusion detection system for industrial IoT networks. Engineering Science and Technology, an International Journal, 38, 101322.

  Kayes, I., & Iamnitchi, A. (2017). Privacy and security in online social networks: A survey. Online Social Networks and Media, 3, 1–21. https://doi.org/10.1016/j.osnem.2017.09.001.

  Kowsari, K., Meimandi, K. J., Heidarysafa, M., Mendu, S., Barnes, L. E., & Brown, D. (2022). Text classification algorithms: A survey. Information, 13(1), 38.

  Mitnick, K. D., & Simon, W. L. (2002). The art of deception: Controlling the human element of security. Wiley Publishing.

  Rovito, L., Bonin, L., Manzoni, L., & De Lorenzo, A. (2022). An evolutionary computation approach for Twitter bot detection. Applied Sciences, 12(12), 5915.

  Salahdine, F., & Kaabouch, N. (2019). Social engineering attacks: A survey. Future Internet, 11(4), 89. https://doi.org/10.3390/fi11040089.

  Sarker, I. H., Ezugwu, A. E., & Chatterjee, J. M. (2021). Cybersecurity data science: An overview from machine learning perspective. Journal of Big Data, 8(1), 1–35.

  Wang, Z., Ren, Y., Zhu, H., & Sun, L. (2022). Threat detection for general social engineering attack using machine learning techniques. arXiv. https://doi.org/10.48550/arXiv.2203.07933.

  Yousef, A. M., Zain, J. M., & Alazab, M. (2023). SIEM-enhanced IPS with deep learning for high-speed networks. Computers & Security, 122, 102920. https://doi.org/10.1016/j.cose.2023.102920.

  Zhou, Y., Qin, Z., Xie, S., & Wang, Y. (2023). A hybrid deep learning model for detecting cyberbullying on social media. IEEE Transactions on Computational Social Systems, 10(2), 345–356.