ImportError: cannot import name 'LatentDirichletAllocation'

问题

I'm trying to import the following:

from sklearn.model_selection import train_test_split

and got following error, here's the stack trace :

ImportError                               Traceback (most recent call last)
<ipython-input-1-bdd2a2f20673> in <module>
      2 import pandas as pd
      3 from sklearn.model_selection import train_test_split
----> 4 from sklearn.tree import DecisionTreeClassifier
      5 from sklearn.metrics import accuracy_score
      6 from sklearn import tree

~/.local/lib/python3.6/site-packages/sklearn/tree/__init__.py in <module>
      4 """
      5 
----> 6 from ._classes import BaseDecisionTree
      7 from ._classes import DecisionTreeClassifier
      8 from ._classes import DecisionTreeRegressor

~/.local/lib/python3.6/site-packages/sklearn/tree/_classes.py in <module>
     38 from ..utils.validation import check_is_fitted
     39 
---> 40 from ._criterion import Criterion
     41 from ._splitter import Splitter
     42 from ._tree import DepthFirstTreeBuilder

~/.local/lib/python3.6/site-packages/sklearn/tree/_splitter.pxd in init sklearn.tree._criterion()

~/.local/lib/python3.6/site-packages/sklearn/tree/_tree.pxd in init sklearn.tree._splitter()

~/.local/lib/python3.6/site-packages/sklearn/neighbors/_quad_tree.pxd in init sklearn.tree._tree()

~/.local/lib/python3.6/site-packages/sklearn/neighbors/__init__.py in <module>
     15 from ._kde import KernelDensity
     16 from ._lof import LocalOutlierFactor
---> 17 from ._nca import NeighborhoodComponentsAnalysis
     18 from ._base import VALID_METRICS, VALID_METRICS_SPARSE
     19 

~/.local/lib/python3.6/site-packages/sklearn/neighbors/_nca.py in <module>
     20 from ..base import BaseEstimator, TransformerMixin
     21 from ..preprocessing import LabelEncoder
---> 22 from ..decomposition import PCA
     23 from ..utils.multiclass import check_classification_targets
     24 from ..utils.random import check_random_state

~/.local/lib/python3.6/site-packages/sklearn/decomposition/__init__.py in <module>
     17 from ._factor_analysis import FactorAnalysis
     18 from ..utils.extmath import randomized_svd
---> 19 from ._online_lda import LatentDirichletAllocation
     20 
     21 __all__ = ['DictionaryLearning',

ImportError: cannot import name 'LatentDirichletAllocation'

Actually I was trying to import :

from sklearn.cross_validation import train_test_split

Which gave me following error:

ModuleNotFoundError                       Traceback (most recent call last)
<ipython-input-2-9ebede864c4d> in <module>
      1 import numpy as np
      2 import pandas as pd
----> 3 from sklearn.cross_validation import train_test_split
      4 from sklearn.tree import DecisionTreeClassifier
      5 from sklearn.metrics import accuracy_score

ModuleNotFoundError: No module named 'sklearn.cross_validation'

So, as answered in this post, I used sklearn.model_selection instead of sklearn.cross_validation. But it's still giving an error.

I'm using python 3.6.8.

回答1:

This is broken in 0.22 version

Use stable versions:

For Python 2

pip uninstall scikit-learn
pip install -U scikit-learn==0.20.4

For Python 3

pip3 uninstall scikit-learn
pip3 install -U scikit-learn==0.21.3

回答2:

This is probably due to an old pip version. You can refer to https://stackoverflow.com/a/59328446/6513708

But long story short, you need to remove the following files which were let by pip during the update:

~/.local/lib/python3.6/site-packages/sklearn/decomposition/_online_lda.cpython-36m-x86_64-linux-gnu.so
~/.local/lib/python3.6/site-packages/sklearn/feature_extraction/_hashing.cpython-36m-x86_64-linux-gnu.so

来源：https://stackoverflow.com/questions/59320664/importerror-cannot-import-name-latentdirichletallocation

标签

python

python-3.x

scikit-learn

sklearn-pandas

train-test-split