mifs安装和sklearn

Question

但是，如果出现sklearn错误，我仍然要安装mif（用于功能选择）：

---------------------------------------------------------------------------
ImportError                               Traceback (most recent call last)
<ipython-input-1-f855ea4d470e> in <module>
     10 from sklearn.ensemble import RandomForestClassifier # random forest for classification
     11 from sklearn.metrics import precision_score # precision
---> 12 import mifs # feature selection
     13 from sklearn.model_selection import RandomizedSearchCV # parameters tuning
     14 import itertools

~/.local/lib/python3.5/site-packages/mifs/__init__.py in <module>
      6 """
      7 
----> 8 from .mifs import MutualInformationFeatureSelector
      9 from .version import _check_module_dependencies, __version__
     10 

~/.local/lib/python3.5/site-packages/mifs/mifs.py in <module>
     10 from sklearn.utils import check_X_y
     11 from sklearn.preprocessing import StandardScaler
---> 12 from sklearn.externals.joblib.parallel import cpu_count
     13 import bottleneck as bn
     14 from . import mi

ImportError: No module named 'sklearn.externals.joblib.parallel'

很显然，我无法通过简单的sklearn.externals.joblib.parallel安装此pip install（我已经尝试过了）。有什么想法如何安装此模块吗？

如果您需要某种澄清，请问。