Synthetic Network Spoofing Detection Dataset: 100,000 Traffic Records for AI-Based Multi-Layer Spoofing Detection

This dataset contains 100,000 synthetic network traffic records generated to evaluate AI-based multi-layer spoofing detection systems. The dataset encompasses five attack categories: IP spoofing, MAC spoofing, ARP spoofing, DNS spoofing, and DDoS attacks, alongside normal traffic (85,000 records). Each record contains 45 engineered features derived from network-level, device-level, and behavioral-level attributes.

Files included:- raw_traffic.csv: Raw simulated network traffic- labeled_dataset.csv: Fully labeled and feature-engineered dataset- train_data.csv: Training split (70,000 records)- validation_data.csv: Validation split (15,000 records)- test_data.csv: Test split (15,000 records)- device_profiles.csv: Per-device behavioral profiles (500 devices)- threat_intelligence.csv: Temporal threat intelligence sequences- cyber.ipynb: Dataset generation and preprocessing notebook

This dataset was used in the paper: "AI-Based Multi-Layer Spoofing Detection System Using Deep Learning and Adaptive Policy Networks for Real-Time Network Security," submitted to Springer Telecommunication Systems Journal, 2026.