More statistics for function thresholder

[BenjaK]

The function thresholder currently computes Sensitivity, Specificity, Youden's J and the distance to the best possible cutoff. However, these may not be the metrics that the user wishes to compare models by.

Internally, thresholder uses caret::confusionMatrix which calculates a bunch of other statistics, but throws away all but those above. This pull request enables the user to choose which statistics to compute for each probability cutoff, the options being the same as those calculated in caret::confusionMatrix. The default option calculates all statistics from caret::confusionMatrix.

[codecov-io]

Codecov Report

Merging #938 into master will decrease coverage by 0.02%.
The diff coverage is n/a.

@@            Coverage Diff             @@
##           master     #938      +/-   ##
==========================================
- Coverage   19.54%   19.51%   -0.03%     
==========================================
  Files          90       90              
  Lines       12077    12093      +16     
==========================================
  Hits         2360     2360              
- Misses       9717     9733      +16

Impacted Files	Coverage Δ
R/thresholder.R	0% <0%> (ø)	⬆️

---

Continue to review full report at Codecov.

Legend - Click here to learn more
Δ = absolute <relative> (impact), ø = not affected, ? = missing data
Powered by Codecov. Last update 464500b...19a10ab. Read the comment docs.

[topepo]

Thanks