Introduction

ImportError: No module named 'sklearn.lda' usually means the code is using an old scikit-learn import path. Linear Discriminant Analysis used to be imported from sklearn.lda in very old versions, but modern scikit-learn moved it to sklearn.discriminant_analysis. The fix is usually just updating the import and, if needed, cleaning up the environment so the installed package version matches the code.

The Old Import Path Is Obsolete

Code like this is the source of the error:

from sklearn.lda import LDA

That module path no longer exists in modern scikit-learn. The current import is:

from sklearn.discriminant_analysis import LinearDiscriminantAnalysis

model = LinearDiscriminantAnalysis()

So the issue is not that scikit-learn is missing LDA. It is that the module layout changed.

Update the Code, Not Just the Package

If your notebook or script came from an old tutorial, the best fix is to modernize the code instead of trying to downgrade the entire library just to preserve the outdated import.

A small working example:

1from sklearn.datasets import load_iris
2from sklearn.discriminant_analysis import LinearDiscriminantAnalysis
3
4X, y = load_iris(return_X_y=True)
5model = LinearDiscriminantAnalysis()
6model.fit(X, y)
7print(model.predict(X[:5]))

This is the modern equivalent of what old sklearn.lda examples were usually trying to do.

Check Which Environment You Are Actually Using

Sometimes the code is already correct, but the environment is not the one you think it is. A notebook kernel, virtual environment, or IDE interpreter may be different from the one where you installed scikit-learn.

Useful checks:

1import sys
2import sklearn
3
4print(sys.executable)
5print(sklearn.__version__)

If the path or version is unexpected, fix the environment before debugging the import further.

LDA Versus QDA in the New Layout

Another source of confusion is that both Linear Discriminant Analysis and Quadratic Discriminant Analysis now live in sklearn.discriminant_analysis.

from sklearn.discriminant_analysis import LinearDiscriminantAnalysis
from sklearn.discriminant_analysis import QuadraticDiscriminantAnalysis

So if older code mentioned sklearn.lda or sklearn.qda, both imports should be updated to the shared modern module.

Do Not Shadow sklearn Locally

Import errors are not always caused by version changes. A local file or directory can shadow the real package.

For example, if your project contains:

sklearn.py

or:

sklearn/

Python may import that local name instead of the installed scikit-learn package. If the behavior seems inconsistent, check the project directory for shadowing names.

A Good Migration Pattern

If you are maintaining older code, replace the old import and rename the class usage clearly.

Old style:

from sklearn.lda import LDA
model = LDA()

Modern style:

from sklearn.discriminant_analysis import LinearDiscriminantAnalysis
model = LinearDiscriminantAnalysis()

This is also more self-explanatory. LinearDiscriminantAnalysis is clearer than the short alias LDA when reading code for the first time.

Common Pitfalls

• Trying to fix the error by reinstalling scikit-learn without changing the obsolete import path.
• Assuming the package is missing LDA entirely. The class still exists, just under a newer module path.
• Forgetting that notebooks and IDEs may use a different Python environment from the shell.
• Confusing LinearDiscriminantAnalysis with QuadraticDiscriminantAnalysis during migration.
• Accidentally shadowing the real sklearn package with a local file or folder of the same name.

Summary

• 'sklearn.lda is an obsolete import path from old scikit-learn code.'
• Use from sklearn.discriminant_analysis import LinearDiscriminantAnalysis instead.
• Check the active Python environment if the import still behaves unexpectedly.
• Update old tutorials and notebooks rather than downgrading your whole stack to preserve a deprecated path.
• Also watch for local files that shadow the installed sklearn package.