Henry Lin
2016-05-21 23:03:17 UTC
Hi, is there any interest in adding the concordance index ranking function
to scikit-learn? It is a ranking score that scores a (finite) ordered list
of objects, and it's used for survival analysis.
Whereas ***@k is concerned about a single object appearing in the top k ranks
of a list, the concordance index is more thorough, judging pairwise scores
of any two objects in the list.
More references:
- Frank Wilcoxon. Individual comparisons by ranking methods
<http://sci2s.ugr.es/keel/pdf/algorithm/articulo/wilcoxon1945.pdf>.
- Vikas C. Raykar et al. On Ranking on Survival Analysis: Bounds on the
Concordance Index
<http://papers.nips.cc/paper/3375-on-ranking-in-survival-analysis-bounds-on-the-concordance-index.pdf>
.
- James C. Costello et al. A community effort to assess and improve drug
sensitivity prediction algorithms
<http://www.nature.com/nbt/journal/v32/n12/full/nbt.2877.html#supplementary-information>
.
to scikit-learn? It is a ranking score that scores a (finite) ordered list
of objects, and it's used for survival analysis.
Whereas ***@k is concerned about a single object appearing in the top k ranks
of a list, the concordance index is more thorough, judging pairwise scores
of any two objects in the list.
More references:
- Frank Wilcoxon. Individual comparisons by ranking methods
<http://sci2s.ugr.es/keel/pdf/algorithm/articulo/wilcoxon1945.pdf>.
- Vikas C. Raykar et al. On Ranking on Survival Analysis: Bounds on the
Concordance Index
<http://papers.nips.cc/paper/3375-on-ranking-in-survival-analysis-bounds-on-the-concordance-index.pdf>
.
- James C. Costello et al. A community effort to assess and improve drug
sensitivity prediction algorithms
<http://www.nature.com/nbt/journal/v32/n12/full/nbt.2877.html#supplementary-information>
.