MultinomialNB Cannot accept negative Alpha

[JohnMerrick]

Hi,

I was trying to run your code on my machine with Phyton 3.7 and at step Multinomial Naive Bayes¶ i get the following error

ValueError Traceback (most recent call last)
in
4 for i in range(-1000,1000,50):
5 clf1 = MultinomialNB(alpha=i)
----> 6 clf1.fit(X_train,y_train)
7 clf1.fit(X_train_2,y_train)
8 scores = cross_val_score(clf1, X_train, y_train, cv=10)

c:\users\johnm\appdata\local\programs\python\python37\lib\site-packages\sklearn\naive_bayes.py in fit(self, X, y, sample_weight)
609 dtype=np.float64)
610 self._count(X, Y)
--> 611 alpha = self._check_alpha()
612 self._update_feature_log_prob(alpha)
613 self._update_class_log_prior(class_prior=class_prior)

c:\users\johnm\appdata\local\programs\python\python37\lib\site-packages\sklearn\naive_bayes.py in check_alpha(self)
471 if np.min(self.alpha) < 0:
472 raise ValueError('Smoothing parameter alpha = %.1e. '
--> 473 'alpha should be > 0.' % np.min(self.alpha))
474 if isinstance(self.alpha, np.ndarray):
475 if not self.alpha.shape[0] == self.feature_count.shape[1]:

ValueError: Smoothing parameter alpha = -1.0e+03. alpha should be > 0.

It would Appear that the MultinomialNB function cannot accept a negative alpha value, how did you manage to run the code with a negative Alpha