CODEWITHSUNDEEP
python-3.xscikit-learnmlp
Geinkehdsk
Geinkehdsk

Reputation: 99

MLP regression, error: TypeError: '<=' not supported between instances of 'numpy.ndarray' and 'int'

I am trying to fit a MLP regression model.

My data looks like

>>> x
array([[-0.10423869, -0.26620437, -0.15806682, ...,  0.14673972,
        -0.08155304, -0.18497986],
       [-0.09611467, -0.25407929, -0.14541038, ...,  0.12948089,
        -0.08478664, -0.18818328],
       [-0.11439996, -0.27334441, -0.19392899, ..., -0.05813874,
        -0.0083624 , -0.1901444 ],
       ...,
       [-0.05907788, -0.19151134, -0.00739118, ..., -0.00207883,
        -0.20569605, -0.09713173],
       [ 0.2577889 , -0.19076356, -0.16640778, ...,  0.19883847,
        -0.19295281,  0.28263902],
       [-0.04733956, -0.33666808, -0.24709939, ..., -0.2130735 ,
        -0.22681055,  0.15976231]])
>>> y
array([[0.],
       [0.],
       [0.],
       ...,
       [0.],
       [0.],
       [0.]])

I want to do 5 cross-validation first to find the best parameters,here is what I wrote for just fit 50 of x_train and y_train, just to check:

mlp = MLPRegressor()
parameter_space = {
    'max_iter': [1000,2000,5000],
    'activation': ['relu'],
    'alpha': [0.0001,0.001,0.01],
    'hidden_layer_sizes': [(8,8,),(50,50,),(100,100,)],
    'solver': ['sgd', 'adam'],
    'learning_rate': ['constant','adaptive']
}
k = 5
kf = KFold(n_splits=k, random_state=None)
model = MLPRegressor(parameter_space)
acc_score = []
for train_index , test_index in kf.split(x):
    x_train , x_valid = x[train_index,:],x[test_index,:]
    y_train , y_valid = y[train_index] , y[test_index]
    model.fit(x_train[1:50], y_train[1:50])

However, it shows this error:

Traceback (most recent call last):
  File "<stdin>", line 4, in <module>
  File "/local/xx/anaconda3/lib/python3.9/site-packages/sklearn/neural_network/_multilayer_perceptron.py", line 673, in fit
    return self._fit(X, y, incremental=False)
  File "/local/xx/anaconda3/lib/python3.9/site-packages/sklearn/neural_network/_multilayer_perceptron.py", line 358, in _fit
    if np.any(np.array(hidden_layer_sizes) <= 0):
TypeError: '<=' not supported between instances of 'numpy.ndarray' and 'int'

Both x_train[1:50] and y_train[1:50] are <class 'numpy.ndarray'>.. How should I fix this?

Thank you

Upvotes: 1

Views: 312

Answers (1)

Kabilan Mohanraj
Kabilan Mohanraj

Reputation: 1916

The parameter_space dictionary can be passed to the estimator to be unpacked. To do so, add 2 stars before the parameter like so **parameter_space when passing. The parameter_space variable has to have the parameters in the supported data types. For example,

parameter_space = {
    'max_iter': 1000,
    'activation': 'relu',
    'alpha': 0.0001,
    'hidden_layer_sizes': (8,50,100),
    'solver': 'sgd',
    'learning_rate': 'constant'
}

model = MLPRegressor(**parameter_space)

The above model gave the desired output without any errors. For more information on MLPRegressor refer here.

Upvotes: 1

Related Questions