Closed koadman closed 12 years ago
After some discussion with Aaron we're going to hold off on this for the time being in favor of more useful and less complex open issues.
pushed with commit 9786306 calculation of recall for taxonomic structure estimation purposes requires special simulated data where all reads should align to markers
The benchmark currently reports "accuracy" at each taxonomic level. It would ideally report precision and recall of placement, along with the F1 score. These metrics could be reported on a per-organism basis as well, and if metadata about the test organism such as its phylogenetic distance from other taxa were available, the relationship between that metadata and accuracy could be shown.