Re: [scikit-learn] sklearn Pipeline: argument of type 'ColumnTransformer' is not iterable

2020-05-29 Thread Chamila Wijayarathna
Thank you both for your inputs. On Fri, May 29, 2020 at 9:57 PM Nicolas Hug wrote: > Also, you should not scale your input before computing cross-validation > scores. By doing that you are biasing your results because each test set > knows something about the rest of the data (even if it's not t

Re: [scikit-learn] sklearn Pipeline: argument of type 'ColumnTransformer' is not iterable

2020-05-29 Thread Nicolas Hug
Also, you should not scale your input before computing cross-validation scores. By doing that you are biasing your results because each test set knows something about the rest of the data (even if it's not target data) The scaling should be applied independently on each (train / test) pair. Th

Re: [scikit-learn] sklearn Pipeline: argument of type 'ColumnTransformer' is not iterable

2020-05-29 Thread Thomas J Fan
Once x = preprocessing.scale(df1) is called, the input to your estimator is no longer a dataframe, so the column transformer can not use strings to select columns. Thomas > On Friday, May 29, 2020 at 11:46 AM, Chamila Wijayarathna > mailto:cdwijayarat...@gmail.com)> wrote: > Hi, > > Thanks, t

Re: [scikit-learn] sklearn Pipeline: argument of type 'ColumnTransformer' is not iterable

2020-05-29 Thread Chamila Wijayarathna
Hi, Thanks, this solution fixed the issue. However, it introduces a new error, which was not there before. Traceback (most recent call last): File "C:\Users\ASUS\PycharmProjects\swelltest\venv\lib\site-packages\sklearn\utils\__init__.py", line 425, in _get_column_indices all_columns = X.col

Re: [scikit-learn] major league hacking summer internship program

2020-05-29 Thread Andreas Mueller
Thanks folks! That gives us a good start I think! Re documentation: honestly I'm not entirely sure if those are good issues because I'm not sure if we have consensus what we want to recommend. We can certainly include these but they require some decisions and a lot of expertise. Maybe we can discu

Re: [scikit-learn] sklearn Pipeline: argument of type 'ColumnTransformer' is not iterable

2020-05-29 Thread Thomas J Fan
VotingClassifer also needs names: ens = VotingClassifier(estimators=[('pipe1', pipe_phy), ('pipe2', pipe_fa)]) Thomas > On Friday, May 29, 2020 at 2:33 AM, Chamila Wijayarathna > mailto:cdwijayarat...@gmail.com)> wrote: > Hi all, > > I did manage to get the code to run using a workaround, which

[scikit-learn] major league hacking summer internship program

2020-05-29 Thread Thomas J Fan
I can commit to reviewing. Diving into their program, it looks like they are hiring supervisers through: https://raise.dev/Apply/?ref=mlh which is titled "Software Developer Coach". By looking at their https://fellowship.mlh.io/students they have about 9 weeks of actual contributing. Given the

Re: [scikit-learn] major league hacking summer internship program

2020-05-29 Thread Guillaume LemaƮtre
Hey, I can dedicate some time to review. Cheers, On Fri, 29 May 2020 at 11:43, Adrin wrote: > Thanks Andy, sounds pretty cool. > > I can commit some reviewing time. There should be maybe two of us at least > that they know they can ping, and we can ping others if needed. > > Cheers, > Adrin >

Re: [scikit-learn] major league hacking summer internship program

2020-05-29 Thread Adrin
Thanks Andy, sounds pretty cool. I can commit some reviewing time. There should be maybe two of us at least that they know they can ping, and we can ping others if needed. Cheers, Adrin On Thu, May 28, 2020 at 11:35 PM Andreas C. Mueller < andreasmuelle...@gmail.com> wrote: > Hi Folks. > > So t