AttributeError: 'str' object has no attribute 'decode' in fitting Logistic Regression Model

Question

I am currently trying to create a binary classification using Logistic regression. Currently I am in determining the feature importance. I already did the data preprocessing (One Hot Encoding and sampling) and ran it with XGBoost and RandomFOrestClassifier, no problem

However, when I tried to fit a LogisticRegression model (below is my code in Notebook),

from sklearn.linear_model import LogisticRegression

#Logistic Regression
# fit the model
model = LogisticRegression()
# fit the model
model.fit(np.array(X_over), np.array(y_over))
# get importance
importance = model.coef_[0]
# summarize feature importance
df_imp = pd.DataFrame({'feature':list(X_over.columns), 'importance':importance})
display(df_imp.sort_values('importance', ascending=False).head(20))

# plot feature importance
plt.bar(list(X_over.columns), importance)
plt.show()

it gave an error

...
~\AppData\Local\Continuum\anaconda3\lib\site-packages\joblib\parallel.py in <listcomp>(.0)
    223         with parallel_backend(self._backend, n_jobs=self._n_jobs):
    224             return [func(*args, **kwargs)
--> 225                     for func, args, kwargs in self.items]
    226 
    227     def __len__(self):

~\AppData\Local\Continuum\anaconda3\lib\site-packages\sklearn\linear_model\_logistic.py in _logistic_regression_path(X, y, pos_class, Cs, fit_intercept, max_iter, tol, verbose, solver, coef, class_weight, dual, penalty, intercept_scaling, multi_class, random_state, check_input, max_squared_sum, sample_weight, l1_ratio)
    762             n_iter_i = _check_optimize_result(
    763                 solver, opt_res, max_iter,
--> 764                 extra_warning_msg=_LOGISTIC_SOLVER_CONVERGENCE_MSG)
    765             w0, loss = opt_res.x, opt_res.fun
    766         elif solver == 'newton-cg':

~\AppData\Local\Continuum\anaconda3\lib\site-packages\sklearn\utils\optimize.py in _check_optimize_result(solver, result, max_iter, extra_warning_msg)
    241                 "    https://scikit-learn.org/stable/modules/"
    242                 "preprocessing.html"
--> 243             ).format(solver, result.status, result.message.decode("latin1"))
    244             if extra_warning_msg is not None:
    245                 warning_msg += "\n" + extra_warning_msg

AttributeError: 'str' object has no attribute 'decode'    

I googled it and mostly all the responses said that this error is because the scikit-learn library tried to decode an already decoded string. But I don't know how to solve it in my case here. I made sure all my data is either integer or float64, and no strings.

Answer 1

I tried to upgrade my scikit-learn using the below command, still, that didn't solve the AttributeError: 'str' object has no attribute 'decode' issue

pip install scikit-learn  -U

Finally, below code snippet solved the issue, add the solver as liblinear

model = LogisticRegression(solver='liblinear')

Answer 2

In the most recent version of scikit-learn (now 0.24.1) the problem has been fixed enclosing a part of code in a try-catch block which I report below: the file is

optimize.py -> _check_optimize_result(solver, result, max_iter=None,
                       extra_warning_msg=None)

and the code piece is

if solver == "lbfgs":
    if result.status != 0:
        try:
            # The message is already decoded in scipy>=1.6.0
            result_message = result.message.decode("latin1")
        except AttributeError:
            result_message = result.message
            warning_msg = (
                "{} failed to converge (status={}):\n{}.\n\n"
                "Increase the number of iterations (max_iter) "
                "or scale the data as shown in:\n"
                "    https://scikit-learn.org/stable/modules/"
                "preprocessing.html"
            ).format(solver, result.status, result_message)

Which was just

if solver == "lbfgs":
    if result.status != 0:
        warning_msg = (
            "{} failed to converge (status={}):\n{}.\n\n"
            "Increase the number of iterations (max_iter) "
            "or scale the data as shown in:\n"
            "    https://scikit-learn.org/stable/modules/"
            "preprocessing.html"
        ).format(solver, result.status, result.message.decode("latin1"))

before. So upgrading scikit-learn solves the problem.

Answer 3

There is a bug with solver='lbfgs'. Changing to 'sag' works around it.