问题
Are there any implementations of One class classifiers in the Spark?
There doesn't appear to be anything in ML or MLlib, but I was hoping that there was an extension developed by someone in the community that would provide some way of producing a trained classification model where only one labeled class is available in the training data.
回答1:
It's Java, not Spark, but LibSVM has a one class SVM classifer, and calling it from Spark shouldn't be a problem.
来源：https://stackoverflow.com/questions/42254637/one-class-classification-models-in-spark