erikerlandson commented on issue #22087: [SPARK-25097][ML] Support prediction on single instance in KMeans/BiKMeans/GMM URL: https://github.com/apache/spark/pull/22087#issuecomment-465365009 This might be too much of a change, but the need for `labelCol` is a training requirement, so if things were factored so that training was decoupled from evaluation (analogous to Estimators & Transformers in ML pipelines?), the evaluation side would be unifiable. Possibly training could be organized so that trainers requiring a label are more specific subclass of unsupervised trainers that do not?
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on GitHub and use the URL above to go to the specific comment. For queries about this service, please contact Infrastructure at: [email protected] With regards, Apache Git Services --------------------------------------------------------------------- To unsubscribe, e-mail: [email protected] For additional commands, e-mail: [email protected]
