RMSLE (root mean squared log error) #20326
Conversation
ogrisel
left a comment
There was a problem hiding this comment.
I think we should also register neg_root_mean_squared_log_error in sklearn.metrics.SCORERS to make it possible to use it with the model selection tools, for instance by passing scoring="neg_root_mean_squared_log_error" to cross_validate.
fixing Indentation error
fixing indentation and removing extra examples
Co-authored-by: Olivier Grisel <[email protected]>
Co-authored-by: Olivier Grisel <[email protected]>
|
I had approved all the proposed changes still it's not working :( how to know that what's going wrong with ci/circle , linting ? |
|
The message says: |
Co-authored-by: Olivier Grisel <[email protected]>
Co-authored-by: Olivier Grisel <[email protected]>
Okay If it will fail this time then I will write the code from the start,,,let's see :) |
No need, it's a better strategy to read the error message to understand what's happening. |
|
For the next step, could you please address: #20326 (review) ? |
Co-authored-by: Olivier Grisel <[email protected]>
Sure! but I need some explanation of this, to understand more about it and the implementation part (How to implement). |
updating contributors docs
|
Hey @ogrisel I got your point and I think we should definitely do that to
Hey @ogrisel , I got your point and I think we should definitely do this to 'cross_validate` |
Co-authored-by: Olivier Grisel <[email protected]>
doc/whats_new/v1.0.rst
Outdated
| ....................... | ||
|
|
||
| - |Feature| :func:`~sklearn.metrics.mean_squared_log_error` now supports | ||
| `squared='False'`. :pr:`20326` by :user:`Uttam kumar <helper-uttam>`_. |
There was a problem hiding this comment.
The doc CI is failing because of those two warnings:
doc/whats_new/v1.0.rst:587: WARNING: Field list ends without a blank line; unexpected unindent.
doc/whats_new/v1.0.rst:589: WARNING: Mismatch: both interpreted text role prefix and reference suffix.
The error messages are weird but I suspect this might be caused because there is already an section named sklearn.metrics upper in the same file. The module sections are ordered by alphabetical order.
There was a problem hiding this comment.
Ops! I thought it was occurring due to some whitespaces or extra lines. :)
There was a problem hiding this comment.
Please move this entry to the sklearn.metrics on line 410 in the same file to avoid redefining the sklearn.metrics section.
There was a problem hiding this comment.
Thanks for the suggestion, I didn't noticed this previously!
I had moved to 410.
Co-authored-by: Olivier Grisel <[email protected]>
|
Again same warning! |
ogrisel
left a comment
There was a problem hiding this comment.
LGTM. Thanks for your contribution @helper-uttam. Once this this merged, please update your main branch and feel free to open a new PR to add scorer I mentioned previously in the discussion.
RMSLE (root mean squared log error) (scikit-learn#20326)
Co-authored-by: Olivier Grisel <[email protected]>
Reference Issues/PRs
Fixes #20318
What does this implement/fix? Explain your changes.
Updating the function
mean_square_log_errorand includeing one more argument in it i.e.,squared=which will be True by default, because in many contests or workplace there is a need of finding root_mean_square_log_error(RMSLE). And scikit-learn initially do not have such inbuilt function but it has 'mean_square_log_error' and we can getroot_mean_square_log_error' by taking the square root ofmean_square_log_error'.Any other comments?