Re: [ML] Tuning the hyper-parameters of an estimator/evaluator

2018-07-04 Thread Yury Babak
Hi, Alexey. Yes, I think we need it. And if you are ready for this task I created it for you: https://issues.apache.org/jira/browse/IGNITE-8924 Regards, Yury -- Sent from: http://apache-ignite-developers.2346864.n4.nabble.com/

[ML] Tuning the hyper-parameters of an estimator/evaluator

2018-07-04 Thread Alexey Zinoviev
Hi, Igniters, I suggest to add analogue of Parameter Grid from scikit-learn to tune hyper-parameters in Cross-Validation process. Currently, the Ignite ML module has only Cross-Validation support and the user needs to find the best parameters manually in a few while-cycles. Yes, I could do this