Python Sklearn - RandomForest and Missing values

Question

I'm trying to perfome RandomForest on a dataset that contain missing values.

My data set looks like :

train_data = [['1' 'NaN' 'NaN' '0.0127034' '0.0435092']
 ['1' 'NaN' 'NaN' '0.0113187' '0.228205']
 ['1' '0.648' '0.248' '0.0142176' '0.202707']
 ..., 
 ['1' '0.357' '0.470' '0.0328121' '0.255039']
 ['1' 'NaN' 'NaN' '0.00311825' '0.0381745']
 ['1' 'NaN' 'NaN' '0.0332604' '0.2857']]

To impute the "NaN" value, I'm using:

from sklearn.preprocessing import Imputer

imp=Imputer(missing_values='NaN',strategy='mean',axis=0)
imp.fit(train_data[0::,1::])
new_train_data=imp.transform(train_data)

But I'm getting the following error:

Traceback (most recent call last):
  File "./RandomForest.py", line 72, in <module>
    new_train_data=imp.transform(train_data)
  File "/home/aurore/.local/lib/python2.7/site-packages/sklearn/preprocessing    /imputation.py", line 388, in transform
    values = np.repeat(valid_statistics, n_missing)
  File "/usr/lib/python2.7/dist-packages/numpy/core/fromnumeric.py", line 343, in repeat
    return repeat(repeats, axis)
ValueError: a.shape[axis] != len(repeats)

I did it:

new_train_data = imp.fit_transform(train_data)

Then I get this error:

Traceback (most recent call last):
  File "./RandomForest.py", line 82, in <module>
    forest = forest.fit(train_data[0::,1::],train_data[0::,0])
  File "/home/aurore/.local/lib/python2.7/site-packages/sklearn/ensemble/forest.py", line 224, in fit
    X, = check_arrays(X, dtype=DTYPE, sparse_format="dense")
  File "/home/aurore/.local/lib/python2.7/site-packages/sklearn/utils/validation.py", line 283, in check_arrays
    _assert_all_finite(array)
  File "/home/aurore/.local/lib/python2.7/site-packages/sklearn/utils/validation.py", line 43, in _assert_all_finite
    " or a value too large for %r." % X.dtype)
 ValueError: Input contains NaN, infinity or a value too large for dtype('float32').

Is there some problem with the package? Can someone please help me? What does it mean?

Answer 1

You train the imputer on columns 1::, but then you try to apply it to all columns. That doesn't work. Do

new_train_data = imp.fit_transform(train_data)