Detection of phishing attacks using machine learning
Siddharth, Rajeev Srivastava, Harsh Raj, Shourya Dwivedi, Shourya Dwivedi, Rohit Singh, Nitish Chaurasiya · 2025
Recent Developments in IOT Technologies (Review) with the increasing traffic in confidential data being passed over networks, it becomes vulnerable to various security threats that can affect its confidentiality, integrity and availability. Intrusion Detection System (IDS) helps to monitor and provide alerts in case of any malicious activity on the network. Though many machine learning methodologies have had their successes in the anomaly detection domain, very few attempt to reflect the sequential nature of network data. The authors of this study explicitly apply a sequential methodology and choose to compare multiple models, including Random Forests (RF) Multi-Layer Perceptrons (MLP), and Long-, Short-Term Memory (LSTM) on the CIDDS-001 dataset. Our evaluations show that the sequential detection is more effective than typical point-wise approaches. In the experimental results, it shows that long short-term memory significantly outperforms in detection orderdependent traffic data logs of tracing code-level anomalies with 99.94% accuracy and an F1-score of 91.66%.