On 8 Sep 2017 5:23 am, "Jimmy Wan" ***@***.***> wrote: Description BaggingRegressor will raise an exception: ValueError: Input contains NaN, infinity or a value too large for dtype('float64'). I propose relaxing this constraint as the wrapped regressor should already handle checking of their input arrays. At best, the check is redundant. At worst, I can't reuse this component. Steps/Code to Reproduce Create pipeline that includes varying methods of imputation and wrap it with a BaggingRegressor and run it on input containing nan/infinity. — You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub <#9708>, or mute the thread <https://github.com/notifications/unsubscribe-auth/AAEz691ELC3kzZ2yTW1lMFSmfve_OjH5ks5sgELBgaJpZM4PQRwX> .