Anomaly output metric

[RobertSellers]

Hi again. Hopefully this is a relatively easy question. What metric (or metrics) would you recommend to quantify the output of each individual track per the A Contrario (or preceding) output? I'm referring to something that could be relatively easily extracted from the code. Would this be the average log probability weight per track? How might a representational value get extracted per MMSI for all tested tracks?

[duonginspace]

Hi,
You could use the NFA to evaluate the "normalcy level" of each track in the test set.

[RobertSellers]

Good advice at making sense of the output, thanks for that.

[RobertSellers]

Hi again -

Any thoughts on the following? Inside the contrario_detection( ) function in contrario_utils.py, what if there is a loop added above the NFA function with additional (more than the default 0.0091) epsilon values and multiple sensitivities approaching 1e-10 to100? That way the totals can be tallied and sorted? Do you think this make sense to you as well?