ML train and test data statistics

stratosphereips / StratosphereLinuxIPS

Slips, a free software behavioral Python intrusion prevention system (IDS/IPS) that uses machine learning to detect malicious behaviors in the network traffic. Stratosphere Laboratory, AIC, FEL, CVUT in Prague.

Other

705 stars 176 forks source link

ML train and test data statistics #831

Open whale-withme opened 4 months ago

whale-withme commented 4 months ago

If I want to see the performance of the system after be trained, how can I get the real label with input pcap file? I can only see logs and it not friendly to me, I want to know how did you test ml module?

eldraco commented 3 months ago

We tested the model with our own large dataset of captures that are labeled in the flow level. See here https://mcfp.felk.cvut.cz/publicDatasets/. If you want to test the performance on your data, you can generate flow, then label them with https://github.com/stratosphereips/netflowlabeler (still needs your knowledge). And then use our comparison tools for the final performance calculation https://github.com/stratosphereips/IDPS-Comparison-Tool