Loading...
Loading...
Contributors:
Dataset URI:
8ws5vb8w2nThe Web2VecPhishingDataset is a collection of four machine-learning-ready artifacts designed for research on phishing website detection and feature-space comparison. The artifacts were generated using Web2VecPhishingPipeline (https://github.com/edytafraszczak/Web2VecPhishingPipeline), which supports data ingestion, URL and domain normalization, label management, feature extraction, integration of external datasets, and feature-space harmonization. The release contains four complementary artifacts: - Web2VecFullSpace – 496,850 rows and 236 columns, including 233 candidate features and 183 numeric machine-learning-ready features. It includes URL, DNS, GEO/network, HTML/HTTP, Open PageRank, SSL/TLS, and WHOIS-derived features. The artifact contains 301,151 legitimate and 195,699 phishing observations. - Web2VecUrlSpace – 5,666,075 rows and 101 columns, including 98 candidate features and 97 numeric features. It contains 999,993 legitimate and 4,666,082 phishing observations. - Web2VecUrlMergedSpace – extends the URL-based representation with compatible records from additional datasets. It contains 6,134,211 rows, including 1,265,082 legitimate and 4,869,129 phishing observations. - Web2VecCommonSpace – an identifier-free harmonized representation based on 18 shared numeric lexical features. It contains 6,753,343 rows, including 1,662,284 legitimate and 5,091,059 phishing observations. Label 0 represents legitimate observations and label 1 represents phishing observations. Labels apply to individual observations and should not be interpreted as indicating that all resources associated with a domain, hosting platform, or service are malicious. The artifacts differ in scale, feature richness, and source composition. To reduce information leakage, observations associated with the same normalized host should remain within the same data partition during model evaluation. Source-held-out, temporal, and independently collected test sets are also recommended. A comprehensive analysis of the dataset will be presented in the accompanying introductory article.
Keyword(s):
Cybersecurity; Machine Learning.
MeSH Term(s):
Grants and Funding:
No Grants or Funding Data Available
Associated Publication(s):
No Associated Publication Data Available
Supplementary Data:
No Supplementary Data Available
Related Dataset(s):
No Related Dataset Data Available
The National Library of Medicine (NLM) provides information about biomedical datasets. Inclusion on an NLM Web page does not imply endorsement of, or agreement with, the contents by NLM or the National Institutes of Health. Learn more: NLM Web Policies (nih.gov)