ValueError: Unknown label type in scikit-learn

Question

I try to generate meta-features, so I follow tutorials and write the following:

clf = tree.DecisionTreeClassifier()

clf.fit(X, y)

But it raises ValueError.

File "/usr/local/lib/python2.7/dist-packages/sklearn/tree/tree.py", line 739, in fit
X_idx_sorted=X_idx_sorted)
File "/usr/local/lib/python2.7/dist-packages/sklearn/tree/tree.py", line 146, in fit
check_classification_targets(y)
File "/usr/local/lib/python2.7/dist-packages/sklearn/utils/multiclass.py", line 172, in check_classification_targets
raise ValueError("Unknown label type: %r" % y_type)
ValueError: Unknown label type: 'unknown'

Why it raises?

The dataset consists of floats and integers, class labels are integers. describe() returns this:

         x1       x2       x3       x4      x5       x6       x7       x8  
count   3500.00  3500.00  3500.00  3500.00  3500.0  3500.00  3500.00  3500.00   
unique   501.00   516.00   572.00   650.00   724.0   779.00   828.00   757.00   
top        0.12     0.79     0.82     0.83     1.9     1.68     1.67     2.03   
freq      23.00    25.00    22.00    18.00    16.0    15.00    13.00    14.00   

         x9      x10  ...        x32      x33      x34      x35      x36  
count   3500.00  3500.00  ...    3500.00  3500.00  3500.00  3500.00  3500.00   
unique   730.00   676.00  ...     496.00   504.00   503.00   505.00   486.00   
top        3.27     3.47  ...       0.01     0.58    -0.27    -0.02     0.26   
freq      15.00    16.00  ...      23.00    24.00    26.00    23.00    24.00   

        x37      x38      x39      x40  class  
count   3500.00  3500.00  3500.00  3500.00   3500  
unique   488.00   490.00   492.00   506.00      3  
top       -0.03    -0.07     0.05    -0.19      1  
freq      23.00    25.00    22.00    24.00   1185

Dataset looks like this:

       x33   x34   x35   x36   x37   x38   x39   x40 class  
0     -0.7  0.51  0.34 -0.13 -0.87  0.56 -0.53  0.29     2  
1     1.12   0.6  0.28  2.17  0.18 -0.09 -1.33     1     1  
2     -0.3 -0.07 -0.99 -0.75  1.11  1.35 -1.63   0.1     0  
3    -0.29 -1.62  0.19 -1.04  0.43 -1.82 -1.14 -0.23     1  
4    -0.78 -0.12 -0.35  0.44  0.31 -0.45 -0.23  0.27     0  
5     0.28  0.61  -0.4 -1.96  1.26 -0.72  2.01  0.95     2  
6     0.07  1.91 -0.15 -0.27   1.9  1.14 -0.05  0.04     0  
7     1.52 -1.52 -0.16 -0.41 -0.48 -0.37   0.8   1.3     2  
8    -0.52 -1.41 -3.49  1.74 -0.37 -0.25 -0.63   0.2     2  
9     0.78  0.09  -0.7  1.12 -0.32 -0.43 -0.34 -1.04     2  
10    0.25  0.29 -0.73 -0.02  2.14  1.49  0.02 -2.16     2  
11   -1.72 -0.09  0.43 -0.33 -1.66 -0.73  1.45  2.11     2  
12   -0.01 -2.63 -1.91  0.59   0.8  0.35  1.58 -0.98     2

Its shape is [3500 rows x 41 columns].

Are you sure that all labels are really integers? Can you provide the data? — languitar, Apr 25 '17 at 10:23
Run `X.describe()`. It will output a list of all your columns and their data types (along with more statistics about each col). Check if any of them is 'unknown'. Do the same for `y`. — TLousky, Apr 25 '17 at 10:26
Then it should not give this error. If your data is not sensitive, can you upload it here, or somewhere else and give the link here, so that we can test the same on our machine. By the way, which version of scikit-learn are you using? — Vivek Kumar, Apr 25 '17 at 12:53

score 8 · Accepted Answer · edited May 23 '17 at 12:02

8

There are two probable problem and solutions:
1. does your data has appropriate dimention? check it by X.shape() to insure your data is in appropriate format, you can also check this question
2. Try to convert your data to float by np.asarray(...,dtype=np.float64), you can also check this question

edited May 23 '17 at 12:02

Community

1
1

answered Apr 25 '17 at 10:31

Masoud

1,343
8
25

2

The second solution helps, but there is `dtype` instead of `dbtype`. – evaleria Apr 25 '17 at 16:17

ValueError: Unknown label type in scikit-learn

1 Answers1