问PyCaret不能从'sklearn.model_selection._search‘中导入名称'_check_param_grid’

EN

嗨，我一直在努力修复这个不断出现的错误。我已经在网上尝试了一切，比如重新安装sklearn到GitHub回购和网站上记录的正确版本。有更多的人无法回忆起，但如果有人有任何建议，请告诉我。

ImportError跟踪(最近一次调用)在225号输入，在() 1从pycaret.clustering导入安装程序导入pycaret 2#，从pycaret.clustering导入设置输入create_model、evaluate_model、predict_model、save_model、load_model ->3

文件~\AppData\Roaming\Python\Python38\site-packages\pycaret\clustering_init_.py:1，in

文件~\AppData\Roaming\Python\Python38\site-packages\pycaret\clustering\oop.py:2，in 1从pycaret.internal.pycaret_experiment.utils导入MLUsecase ->2来自pycaret.internal.pycaret_experiment.unsupervised_experiment导入(3 _UnsupervisedExperiment，4)5导入pycaret.internal.patches.sklearn 6导入pycaret.internal.patches.yellowbrick

将~\AppData\Roaming\Python\Python38\site-packages\pycaret\internal\pycaret_experiment\unsupervised_experiment.py:20，文件从pycaret.internal.preprocess.preprocessor导入预处理器19从pycaret.internal.pycaret_experiment.utils导入highlight_setup，MLUsecase -> 20从pycaret.internal.pycaret_experiment.utils导入_TabularExperiment 21从pycaret.internal.pipeline导入( 22管道为InternalPipeline，23 estimator_pipeline，24 get_pipeline_fit_kwargs，25 ) 26从pycaret.internal.utils导入to_df，infer_ml_usecase，mlflow_remove_bad_chars

文件~\AppData\Roaming\Python\Python38\site-packages\pycaret\internal\pycaret_experiment\tabular_experiment.py:18，在16导入pycaret.internal.patches.sklearn 17导入pycaret.internal.patches.yellowbrick --> 18导入pycaret.internal.persistence 19导入pycaret.internal.preprocess 20导入拼图为skplt # type:忽略

从~\AppData\Roaming\Python\Python38\site-packages\pycaret\internal\persistence.py:9，导入管道7从pycaret.internal.utils导入get_logger ->9从pycaret.internal.pycaret_experiment.utils导入add_model_to_pipeline 10从pycaret.internal.pycaret_experiment.utils导入MLUsecase 13 def deploy_model( 14种模型，model_name: str，身份验证:def，platform: str = "aws"，prep_pipe=None 15 )：

文件~\AppData\Roaming\Python\Python38\site-packages\pycaret\utils\time_series\forecasting\pipeline.py:3，in 1从复制导入深拷贝2从sktime.forecasting.base导入BaseForecaster ->3从sktime.forecasting.compose导入ForecastingPipeline 4从sktime.forecasting.compose导入TransformedTargetForecaster 7类

文件~\AppData\Roaming\Python\Python38\site-packages\sktime\forecasting\compose_init_.py:28，in 8 all =9 " ColumnEnsembleForecaster "，10 "EnsembleForecaster"，(.) 24 "make_reduction"，25 27来自sktime.forecasting.compose._column_ensemble import ColumnEnsembleForecaster--> 28来自sktime.forecasting.compose._ensemble import ( 29 AutoEnsembleForecaster，30 EnsembleForecaster，31 ) 32从sktime.forecasting.compose._multiplexer进口MultiplexForecaster 33从sktime.forecasting.compose._pipeline进口( 34 ForecastingPipeline，35 TransformedTargetForecaster，36 )

从sktime.forecasting.base._base导入DEFAULT_ALPHA 20从sktime.forecasting.base._meta _HeterogenousEnsembleForecaster -> 21从sktime.forecasting.model_selection导入temporal_train_test_split 22从sktime.utils.stats导入( 23 _weighted_geometric_mean，24 _weighted_max，25 _weighted_median，26 _weighted_min，27 ) 28来自sktime.utils.validation.forecasting import check_regressor

文件~\AppData\Roaming\Python\Python38\site-packages\sktime\forecasting\model_selection_init_.py:22，in 20从sktime.forecasting.model_selection._split导入SlidingWindowSplitter 21从sktime.forecasting.model_selection._split导入temporal_train_test_split -> 22从sktime.forecasting.model_selection._tune导入ForecastingGridSearchCV 23从sktime.forecasting.model_selection._tune导入ForecastingRandomizedSearchCV

文件~\AppData\Roaming\Python\Python38\site-packages\sktime\forecasting\model_selection_tune.py:13，in 11从sklearn.base导入克隆12从sklearn.model_selection导入ParameterGrid ParameterSampler check_cv -> 13从sklearn.model_selection._search导入_check_param_grid 14从sklearn.utils.metaestimators导入if_delegate_has_method 16从sktime.exceptions import NotFittedError

ImportError:无法从'sklearn.model_selection._search‘导入名称'_check_param_grid’(C:\Users\

hashi\Anaconda3\lib\site-packages\sklearn\model_selection_search.py)