Closed Christtella closed 4 months ago
Thanks @Christtella for this addition :rocket:
@ZimmerD, could you please take a look at this feature addition? I'll take a look at it too, but I know you've contributed here.
@ZimmerD @Christtella, would make sense to couple both input sequences? If you accidentally change one or both sequences during processing (e.g. sorting), the resulting NMI will be corrupted. If the input were an array of tuples instead of two separate sequences, it might be much safer to use. The only drawback I can think of is that the parameter order is no longer intuitive and must be checked from the function description.
//current
let calcNMI (expected: int[]) (actual: int []) = ...
//proposed
let calcNMI (input: (int*int)[]) = ...
Attention: 8 lines
in your changes are missing coverage. Please review.
Comparison is base (
6c97a2b
) 47.16% compared to head (f2b61fb
) 47.31%.
Files | Patch % | Lines |
---|---|---|
src/FSharp.Stats/ML/Unsupervised/ClusterNumber.fs | 77.77% | 0 Missing and 8 partials :warning: |
:umbrella: View full report in Codecov by Sentry.
:loudspeaker: Have feedback on the report? Share it here.
closes #313
Please list the changes introduced in this PR
Description NMI is a measure used to evaluate clustering quality.
[Required] please make sure you checked that
[Optional]