Context Dataset for Microsoft Malware Prediction Kaggle Competition (https://www.kaggle.com/c/microsoft-malware-prediction) Based on work from Chris Deotte & Aditya Soni, this dataset extends previous AvSigVersion timestamps, with detailed added/updated threat lists for each AvSigVersion. Details are scrapped from this page: https://www.microsoft.com/en-us/wdsi/definitions/antimalware-definition-release-notes (See sections New definitions/Updated Definitions) Content - AvSigVersion, Name (of Threat), Alert Level, Type (Added or Updated), ~1M rows Acknowledgements External (Scrapped Datasets) Useful for EDA :) by Aditya Soni https://www.kaggle.com/c/microsoft-malware-prediction/discussion/78672 AvSigVersion and CensusOsVersion TimeStamps by Chris Deotte https://www.kaggle.com/cdeotte/malware-timestamps Notes - Did my best but cannot ensure the file is without errors. As is, do your checks! - Very preliminary analysis didn't spot any clear sign of correlation with high detection spike periods (which is strange?), needs more work. (ex: name column can be further transformed for extended analysis) - 25 September is a strange day for example, would say it has some meaning on the splits (last day for train?), and a huge list of threats added/updated (~7k rows?) - ex: https://www.microsoft.com/en-us/wdsi/definitions/antimalware-definition-release-notes?RequestVersion=1.275.1733.0