For mono-task outputs, use %s

Package:

scikit-learn

47032

Exception Class:

ValueError

Raise code

        y = y.astype(X.dtype)

        if hasattr(self, 'l1_ratio'):
            model_str = 'ElasticNet'
        else:
            model_str = 'Lasso'
        if y.ndim == 1:
            raise ValueError("For mono-task outputs, use %s" % model_str)

        n_samples, n_features = X.shape
        _, n_tasks = y.shape

        if n_samples != y.shape[0]:
            raise ValueError("X and y have inconsistent dimensions (%d != %d)"
                             % (n_samples, y.shape[0]))

Links to the raise (1)

https://github.com/scikit-learn/scikit-learn/blob/c67518350f91072f9d37ed09c5ef7edf555b6cf6/sklearn/linear_model/_coordinate_descent.py#L1920

Ways to fix

If the shape of the label array is 1D then the model ElasticNet should be used instead of MultiTaskElasticNet.

Reproducing the error:

from sklearn import linear_model
clf = linear_model.MultiTaskElasticNet(alpha=0.1,)
X = [[0,0], [1, 1], [2, 2]]
y = [[0, 0], [1, 1], [2, 2]]
y = [0,1,1]
clf.fit(X,y)
print(clf.coef_)

Output Error:

---------------------------------------------------------------------------
ValueError                                Traceback (most recent call last)
<ipython-input-29-5aa9de9a397b> in <module>()
      4 y = [[0, 0], [1, 1], [2, 2]]
      5 y = [0,1,1]
----> 6 clf.fit(X,y)
      7 print(clf.coef_)

/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_coordinate_descent.py in fit(self, X, y)
   1760             model_str = 'Lasso'
   1761         if y.ndim == 1:
-> 1762             raise ValueError("For mono-task outputs, use %s" % model_str)
   1763 
   1764         n_samples, n_features = X.shape

ValueError: For mono-task outputs, use ElasticNet

Fixed version of the code:

from sklearn import linear_model
clf = linear_model.ElasticNet(alpha=0.1,)
X = [[0,0], [1, 1], [2, 2]]
y = [[0, 0], [1, 1], [2, 2]]
y = [0,1,1]
clf.fit(X,y)
print(clf.coef_)

Output:

[0.20493848 0.20470839]

Jul 15, 2021

kellemnegasi 31.6k

Summary:

This exception is thrown when the fit function is called on an instance of MultiTaskElasticNet. The fit function takes in 2 parameters: X and y. Both of them must be 2 dimensional array-like values. X should be of shape (n_smaples, n_features) and y should be of shape (n_samples, n_tasks). This exception is thrown if the value of y is only one-dimensional. Therefore you should ensure that y is a 2D array to avoid this exception.

Code to Reproduce the Exception (Wrong):

from sklearn.linear_model._coordinate_descent import MultiTaskElasticNet
import numpy as np

mten = MultiTaskElasticNet()
X = np.array([[1,2],[3,4]])
y = np.array([1,2, 3, 4])
mten.fit(X, y)

Error Message:

ValueError                                Traceback (most recent call last)
<ipython-input-46-ab7580b5d789> in <module>()
      5 X = np.array([[1,2],[3,4]])
      6 y = np.array([1,2, 3, 4])
----> 7 mten.fit(X, y)

/usr/local/lib/python3.7/dist-packages/sklearn/linear_model/_coordinate_descent.py in fit(self, X, y)
   1760             model_str = 'Lasso'
   1761         if y.ndim == 1:
-> 1762             raise ValueError("For mono-task outputs, use %s" % model_str)
   1763 
   1764         n_samples, n_features = X.shape

ValueError: For mono-task outputs, use ElasticNet

Working Version (Right):

from sklearn.linear_model._coordinate_descent import MultiTaskElasticNet
import numpy as np

mten = MultiTaskElasticNet()
X = np.array([[1,2],[3,4]])
y = np.array([[1,2],[3,4]])
mten.fit(X, y)

Jul 15, 2021

codingcomedyig 7.9k

Add a possible fix

Please authorize to post fix