predictions_roc

Container name: vorontsovie/predictions_roc

Benchmark to compute basic metrics of binary classification problem given scores of objects in positive and negative classes. Origin of data doesn't matter, but scores should be supplied from outside.

Instructions

Actual labels should be mounted as /benchmark_specific_data/ground_truth.txt:

+
-
-
...

Prediction scores should be supplied in /workdir/config.json with the same objects ordering as labels have. config.json format: {"predictions": [1.23, -3.4, 0.56, ...]}

Results are to be stored in /workdir/persistent/result.json:

{
  "metrics": {
    "roc_auc": roc_auc,
  },
  "supplementary": {
    "roc": [{"fpr": , "tpr": }, {"fpr": , "tpr": },...],
  },
}