Github user mengxr commented on the pull request:
https://github.com/apache/spark/pull/1270#issuecomment-54865164
@avulanov Sorry for getting back late! For the implementation, shall we
define an aggregator and then compute all necessary information in a single
pass, instead of trigger a job for each?
For the metric names, I think our reference is [Mining Multi-label
Data](http://lkm.fri.uni-lj.si/xaigor/slo/pedagosko/dr-ui/tsoumakas09-dmkdh.pdf)
and we should follow the naming there:
1. `strictAccuracy` -> `subsetAccuracy`
2. `microPrecisionDoc` -> `microPrecision` (and update other metric names)
3. add `precision`, `recall`, `fMeasure` and `accuracy` (example-based)
4. For per-class metrics, I suggest removing`Class` and overload the metric
method, as in `MulticlassMetrics`.
---
If your project is set up for it, you can reply to this email and have your
reply appear on GitHub as well. If your project does not have this feature
enabled and wishes so, or if the feature is enabled but not working, please
contact infrastructure at [email protected] or file a JIRA ticket
with INFRA.
---
---------------------------------------------------------------------
To unsubscribe, e-mail: [email protected]
For additional commands, e-mail: [email protected]
