Anomalous Activity Detection

  • This was a personal initiative to read Windows Server event log data and build a profile for access on that machine to be used for identifying what is non-standard behavior. The model was a regression model rather than a classifier, as this would allow the model to essentially come up with its own conclusions in the face of new "routines" as opposed to providing predefined sets of "strange" behavior.
Anomalous Activity Detection

Relevant Skills

PythonUsed for data processing and programming machine learning models
Windows PowerShellUsed for extracting event log data on a schedule
Machine LearningUsed to examine existing event log data and find outliers