EdgePhish-5G Dataset: Phishing URL Detection with 5G Slice Annotation

This dataset contains 340,000 URLs for phishing detection research and 5G-inspired traffic slice annotation. Data sources include PhishTank, OpenPhish, a Kaggle phishing dataset, Alexa Top Sites, and Common Crawl. The dataset is balanced between phishing and legitimate URLs: - Phishing URLs: 170,000- Legitimate URLs: 170,000 Files included in this record:1. balanced_urls.csv —the main dataset of 340,000 labeled URLs 2. urls_annotated.csv — the annotated version including 5G-inspired slice labels and score columns3. urls_dataset.csv — a balanced subset provided for controlled experiments and comparative evaluation Each URL is labeled as phishing or legitimate and may additionally be annotated using rule-based heuristics into one of three 5G-inspired traffic slices:- eMBB- mMTC- URLLC These slice labels are heuristic proxies based on structural and semantic URL characteristics, and do not represent ground-truth operator-assigned 5G network slices. The annotated dataset includes the following columns:- url- label- slice- URLLC_score- mMTC_score This dataset is intended for research in phishing URL detection, cybersecurity, machine learning, edge intelligence, and 5G network security.