Code used during 303COM project to analyse unsupervised anomaly detection algorithms on Splunk and Elastic.
Three key files were used during development of this project. All files pertaining to either Elastic (es) or Splunk (sp) begin with es/sp.
es_add_integers.py- Add all necessary string-to-integer conversions to raw OpTC file for use in Elastic X-pack outlier detection; also adds ground_truth field to help extract results of model
sp_ml_statistics_calc.py- contains calculation class to determine performance of Splunk MLTK anomaly detection algorithms. Requires a CSV file in the format:
_time, pid, hostname, outlier_prediction
utils.py- contains necessary data to determine whether an event is a true positive and key dictionaries used for string-to-integer conversions.