Problems Importing scipy/sklearn in Jupyter Notebook

573 Views Asked by At

So I've been working on machine learning with Random Forests in Jupyter Notebooks:

from sklearn.cross_validation import train_test_split
train, test = train_test_split(MLDF, test_size = 0.2) #split the data 80/20

#No output, but the cell ran fine

model.fit(train[X], train[Y])
predicted= model.predict(test[X])

#Was running, did not finish prior to dropped connection

I was letting my RF code run for about 1 hour, and came back. Unfortunately, I had forgotten to stop my computer from falling asleep, which obviously caused the connection to drop.

But when I tried to run it again, sklearn had stopped working. Here's the error:

ImportError: cannot import name __check_build

When I tried playing around with my first cell, where I did most of my imports, I also got this error:

ImportError: No module named version

After browsing this site, I gathered that I'm either not running the correct version of scipy or scikit-learn or not importing them properly. So I went to my terminal and made sure everything was updated, which it was ("requirement already satisfied"). My question isn't simply what am I doing wrong, but how does it make sense for sklearn to stop working just because of a dropped connection? Yes, I already tried uninstalling and re-installing scipy. Any feedback as to what I am missing is greatly appreciated, thank you.

0

There are 0 best solutions below