search

openmainframeproject / ade

ADE detects anomalous time slices and messages in Linux logs (either RFC3164 or RFC5424 format) using statistical learning.
https://www.openmainframeproject.org/projects/anomaly-detection-engine-for-linux-logs-ade
GNU General Public License v3.0
51 stars 34 forks source link

JOSS review: Performance #89

Open mdpiper opened 3 years ago

mdpiper commented 3 years ago

This issue is raised in reference to JOSS submission 3052.

There's an unsubstantiated performance claim on line 48 of the manuscript. The claim should be backed up with data or struck.

mdpiper commented 3 years ago

A suggestion: A quick comparison between ADE and the other techniques might be all that's necessary. E.g.,

We ran ADE, along with approach A and B, over N time slices of a dataset. ADE outperformed the other approaches by a factor of X.

ayush-1506 commented 3 years ago

This definitely looks like something we should add, thanks for raising this! I'll add some benchmarks (performance and speed) on ADE and another approach.