def get_class_votes(self, X, ht):
dist = {}
prediction_option = ht.leaf_prediction
if prediction_option == MAJORITY_CLASS: #MC
dist = self.get_observed_class_distribution()
elif prediction_option == NAIVE_BAYES: #NB
dist = do_naive_bayes_prediction(
X, self._observed_class_distribution,
self._attribute_observers)
# NBAdaptive
if self._mc_correct_weight > self._nb_correct_weight:
dist = self.get_observed_class_distribution()
else:
dist = do_naive_bayes_prediction(
X, self._observed_class_distribution,
self._attribute_observers)
dist_sum = sum(dist.values()) # sum all values in dictionary
if dist_sum * self.get_error_estimation(
) * self.get_error_estimation() > 0.0:
normalize_values_in_dict(
dist_sum * self.get_error_estimation() *
self.get_error_estimation(), dist)
return dist
def get_class_votes(self, X, ht):
if self.get_weight_seen() >= ht.nb_threshold:
return do_naive_bayes_prediction(
X, self._observed_class_distribution,
self._attribute_observers)
else:
return super().get_class_votes(X, ht)
def learn_from_instance(self, X, y, weight, ht):
if self._observed_class_distribution == {}:
# All classes equal, default to class 0
if 0 == y:
self._mc_correct_weight += weight
elif max(self._observed_class_distribution, key=self._observed_class_distribution.get) == y:
self._mc_correct_weight += weight
nb_prediction = do_naive_bayes_prediction(
X, self._observed_class_distribution, self._attribute_observers)
if max(nb_prediction, key=nb_prediction.get) == y:
self._nb_correct_weight += weight
super().learn_from_instance(X, y, weight, ht)
def get_class_votes(self, X, ht):
if self._mc_correct_weight > self._nb_correct_weight:
return self._observed_class_distribution
return do_naive_bayes_prediction(X,
self._observed_class_distribution,
self._attribute_observers)