zhengruifeng commented on issue #27944: [SPARK-31180][ML] Implement PowerTransform URL: https://github.com/apache/spark/pull/27944#issuecomment-600470611 differences from sklearn's implementation: 1, may group distinct values; 2, ignore standardation, which is enabled in sklearn by default; we can standardize the data by `StandardScaler` 3, both use brent solver, but the implementation may differ; sklearn use `scipy.optimize.brent` with bound=[-2,2], iter=500, tol=1.48e-8, and `scipy.optimize.brent` do not guarantee solution between input bounds; while I use `org.apache.commons.math3.optim.univariate.BrentOptimizer` with bound=[-10,10], iter=1000; (I also tried sklearn's default parameters, the results are the same; I conservatively changed bound and iter)
---------------------------------------------------------------- This is an automated message from the Apache Git Service. To respond to the message, please log on to 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]
