Issue #19614: Warn on irrelevant parameters for linear kernel in SVC #28641

Issac-Kondreddy · Mar 17, 2024

Fixes #XXXX

This pull request addresses the issue raised in #19614, where SVC does not warn users when irrelevant parameters (gamma, coef0, degree) are set for the linear kernel. Such parameters have no effect and can mislead users, especially those new to SVM or scikit-learn.

Changes Made:

Modified svm/_classes.py to issue warnings when gamma, coef0, or degree is set to non-default values while using a linear kernel.

Rationale:

The changes improve user feedback and contribute to a more intuitive user experience by notifying users of potentially unintended parameter configurations. This approach follows scikit-learn's principles of transparency and user-friendliness in machine learning model configuration.

…rnel in SVC

github-actions · Mar 17, 2024

❌ Linting issues

This PR is introducing linting issues. Here's a summary of the issues. Note that you can avoid having linting issues by enabling pre-commit hooks. Instructions to enable them can be found here.

You can see the details of the linting issues under the lint job here

`black`

black detected issues. Please run black . locally and push the changes. Here you can see the detected issues. Note that running black might also fix some of the issues which might be detected by ruff. Note that the installed black version is black=23.3.0.


--- /home/runner/work/scikit-learn/scikit-learn/sklearn/svm/_classes.py	2024-03-17 02:01:18.825534 +0000
+++ /home/runner/work/scikit-learn/scikit-learn/sklearn/svm/_classes.py	2024-03-17 02:01:41.980470 +0000
@@ -319,23 +319,35 @@
         )
 
         check_classification_targets(y)
 
         self.classes_ = np.unique(y)
-        if self.kernel == 'linear':
-            if self.gamma not in ('scale', 'auto'):
+        if self.kernel == "linear":
+            if self.gamma not in ("scale", "auto"):
                 warn(
-                    "The 'gamma' parameter has been set but is not relevant for the 'linear' kernel. It will be ignored.",
-                    UserWarning)
+                    (
+                        "The 'gamma' parameter has been set but is not relevant for the"
+                        " 'linear' kernel. It will be ignored."
+                    ),
+                    UserWarning,
+                )
             if self.coef0 != 0:
                 warn(
-                    "The 'coef0' parameter has been set but is not relevant for the 'linear' kernel. It will be ignored.",
-                    UserWarning)
+                    (
+                        "The 'coef0' parameter has been set but is not relevant for the"
+                        " 'linear' kernel. It will be ignored."
+                    ),
+                    UserWarning,
+                )
             if self.degree != 3:
                 warn(
-                    "The 'degree' parameter has been set but is not relevant for the 'linear' kernel. It will be ignored.",
-                    UserWarning)
+                    (
+                        "The 'degree' parameter has been set but is not relevant for"
+                        " the 'linear' kernel. It will be ignored."
+                    ),
+                    UserWarning,
+                )
 
         _dual = _validate_dual_parameter(
             self.dual, self.loss, self.penalty, self.multi_class, X
         )
 
would reformat /home/runner/work/scikit-learn/scikit-learn/sklearn/svm/_classes.py

Oh no! 💥 💔 💥
1 file would be reformatted, 915 files would be left unchanged.

`ruff`

ruff detected issues. Please run ruff --fix --show-source . locally, fix the remaining issues, and push the changes. Here you can see the detected issues. Note that the installed ruff version is ruff=0.3.3.


warning: The `--show-source` argument is deprecated. Use `--output-format=full` instead.
sklearn/svm/_classes.py:1:1: I001 [*] Import block is un-sorted or un-formatted
   |
 1 | / import warnings
 2 | | from warnings import warn
 3 | | from numbers import Integral, Real
 4 | | 
 5 | | import numpy as np
 6 | | 
 7 | | from ..base import BaseEstimator, OutlierMixin, RegressorMixin, _fit_context
 8 | | from ..linear_model._base import LinearClassifierMixin, LinearModel, SparseCoefMixin
 9 | | from ..utils._param_validation import Hidden, Interval, StrOptions
10 | | from ..utils.multiclass import check_classification_targets
11 | | from ..utils.validation import _num_samples
12 | | from ._base import BaseLibSVM, BaseSVC, _fit_liblinear, _get_liblinear_solver_type
13 | | 
14 | | 
15 | | def _validate_dual_parameter(dual, loss, penalty, multi_class, X):
   | |_^ I001
16 |       """Helper function to assign the value of dual parameter."""
17 |       if dual == "auto":
   |
   = help: Organize imports

sklearn/svm/_classes.py:327:89: E501 Line too long (122 > 88)
    |
325 |             if self.gamma not in ('scale', 'auto'):
326 |                 warn(
327 |                     "The 'gamma' parameter has been set but is not relevant for the 'linear' kernel. It will be ignored.",
    |                                                                                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ E501
328 |                     UserWarning)
329 |             if self.coef0 != 0:
    |

sklearn/svm/_classes.py:331:89: E501 Line too long (122 > 88)
    |
329 |             if self.coef0 != 0:
330 |                 warn(
331 |                     "The 'coef0' parameter has been set but is not relevant for the 'linear' kernel. It will be ignored.",
    |                                                                                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ E501
332 |                     UserWarning)
333 |             if self.degree != 3:
    |

sklearn/svm/_classes.py:335:89: E501 Line too long (123 > 88)
    |
333 |             if self.degree != 3:
334 |                 warn(
335 |                     "The 'degree' parameter has been set but is not relevant for the 'linear' kernel. It will be ignored.",
    |                                                                                         ^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^^ E501
336 |                     UserWarning)
    |

Found 4 errors.
[*] 1 fixable with the `--fix` option.

_{Generated for commit: 4a78b6a. Link to the linter CI: here}

jeremiedbb

Thanks for the PR @Issac-Kondreddy. There are linting issues that you can fix following the instructions in #28641 (comment).

jeremiedbb · Mar 18, 2024

sklearn/svm/_classes.py

        self.classes_ = np.unique(y)
+        if self.kernel == 'linear':
+            if self.gamma not in ('scale', 'auto'):


Let's warn for all non-default values

Suggested change

if self.gamma not in ('scale', 'auto'):

if self.gamma != "scale":

miguelcsilva · Mar 31, 2024

@Issac-Kondreddy let me know if you're still working on this. Otherwise would be happy to take this on.

StefanieSenger · Apr 19, 2024

This PR will also close #19630; in fact it's a bit of a duplicate to it.

lamdang2k · Oct 29, 2024

I assume that this PR is stalled and proposed a new PR to fix linting issues

Issue scikit-learn#19614: Warn on irrelevant parameters for linear ke…

4a78b6a

…rnel in SVC

github-actions bot added the module:svm label Mar 17, 2024

jeremiedbb reviewed Mar 18, 2024

View reviewed changes

lamdang2k mentioned this pull request Oct 29, 2024

ENH Warn on irrelevant parameters for linear kernel in SVC #30174

Open

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

Uh oh!

Uh oh!

Issue #19614: Warn on irrelevant parameters for linear kernel in SVC #28641

Issue #19614: Warn on irrelevant parameters for linear kernel in SVC #28641

Uh oh!

Issac-Kondreddy commented Mar 17, 2024

Uh oh!

github-actions bot commented Mar 17, 2024

Uh oh!

jeremiedbb left a comment

Uh oh!

jeremiedbb Mar 18, 2024

Uh oh!

miguelcsilva commented Mar 31, 2024

Uh oh!

StefanieSenger commented Apr 19, 2024

Uh oh!

lamdang2k commented Oct 29, 2024

Uh oh!

Uh oh!

	if self.gamma not in ('scale', 'auto'):
	if self.gamma != "scale":

Search code, repositories, users, issues, pull requests...

Uh oh!

Issue #19614: Warn on irrelevant parameters for linear kernel in SVC #28641

Are you sure you want to change the base?

Issue #19614: Warn on irrelevant parameters for linear kernel in SVC #28641

Uh oh!

Conversation

Issac-Kondreddy commented Mar 17, 2024

Fixes #XXXX

Changes Made:

Rationale:

Uh oh!

github-actions bot commented Mar 17, 2024

❌ Linting issues

black

ruff

Uh oh!

jeremiedbb left a comment

Choose a reason for hiding this comment

Uh oh!

jeremiedbb Mar 18, 2024

Choose a reason for hiding this comment

Uh oh!

miguelcsilva commented Mar 31, 2024

Uh oh!

StefanieSenger commented Apr 19, 2024

Uh oh!

lamdang2k commented Oct 29, 2024

Uh oh!

Uh oh!

`black`

`ruff`