MeAJOR: Merged email Assets from Joint Open-source Repositories

The MeAJOR dataset was created with multi-source legitimate emails and malicious emails from different attack campaigns to enable a more reliable phishing detection. MeAJOR provides a total of 108685 data samples, in preprocessed format with engineered features suitable for the training, validation, and testing of ML and DL models. The dataset was created at GECAD, ISEP, Polytechnic of Porto. It contains data collected from publicly available spam email and phishing email datasets: TREC-05 TREC-06 TREC-07 Nazario Phishing Corpus Nigerian Fraud The dataset contains the email messages anonymized, source dataset and label (0 = benign; 1 = phishing). It also contains several extracted features suitable for email phishing attack detection. The email messages are anonymized using Entity Removal, with tokens replacing them, enclosed between square brackets. The tokens are as follow: [PGP], [EMOJI], [SYMBOL], [NAME], [USERNAME], [INITIALS], [EMAIL_ADDRESS], [PHONE_NUMBER], [ADDRESS], [ORGANIZATION], [URL], [IP_ADDRESS], [FILE_PATH], [FILE_NAME], [FILE], [DATE], [TIME], [FINANCIAL_INFO], [PRODUCT], [REFERENCE_NUMBER]. The emails are available in both an Apache Parquet file and a comma-separated values (CSV) file. We recommend using the Apache Parquet files for more efficient querying and compression.