gold-miner-auditor: predicts what gold-miner may be good at identifyingΒΆ

*NOTE: this tool is part of the gold-miner-ui python package, which is distributed separately from the gold-miner python package``

The gold-miner approach will work better at differentiating between some protocols vs others. gold-miner-auditor, when given a training profile, attempts to predict which protocols will (hopefully) be better predicted when compared with other traffic types. To determine this, use gold-miner-auditor to display, given a list of traffic labels of interest. The result will be a list of other protocols and the similarity to them (with lower numbers being more similar and less likely to be differentiated against, potentially producing false positives). Use the -j switch to output JSON formatted results.

Example:

$ gold-miner-auditor -d mail -j training-profile.fsdb

Similarity distance for mail from other types (lower is better):

  mail:       0.0
  web:        0.83