NetPhish

A brand-new, well-balanced dataset called NetPhish was produced using Netcraft. it has 50,406 instances and 30 features in total. No duplicates or missing values exist. It also has a helpful binary feature that comes from ownership of the infrastructure. As a result, the dataset is a trustworthy and clean source for phishing detection model training.本研究基于Netcraft生成了一款全新且分布均衡的数据集NetPhish。该数据集总计包含50406条样本与30个特征，不存在重复样本与缺失值。此外，数据集还包含一项源自基础设施所有权的实用二值特征（binary feature）。因此，该数据集可作为训练网络钓鱼检测模型（phishing detection model）的可靠且洁净的数据源。