See the Spark ML Documentation https://spark.apache.org/docs/1.6.1/api/scala/index.html#org.apache.spark.ml.evaluation.BinaryClassificationEvaluator

ml_binary_classification_eval(predicted_tbl_spark, label, score,
  metric = "areaUnderROC")

Arguments

predicted_tbl_spark
The result of running sdf_predict
label
Name of column string specifying which column contains the true, indexed labels (ie 0 / 1)
score
Name of column contains the scored probability of a success (ie 1)
metric
The classification metric - one of: areaUnderRoc (default) or areaUnderPR

Value

area under the specified curve