Atoposs/SpamRegistration

--- license: mit --- **Spam Registration Dataset**: This dataset targets the detection of spam user registrations. It contains over 1.21M accounts collected across seven consecutive days, with a consistently high positive rate (global average 63.32\%). The ratio of positive to negative samples ranges from 1.41 to 2.03, indicating a positive-dominant and relatively stable class distribution over time. Such characteristics reflect scenarios where spam registrations are prevalent and require high-recall detection under purity constraints. - The attribute values in each record are separated by `@!!!!@`. - All entity values have been anonymized to protect user privacy and commercial confidentiality. The domain information of each attribute column is illustrated in the figure below：