Warn against dodgy classification

Author: bricef

06-Naive-Bayes/general_classifier.py

@@ -25,6 +25,9 @@ def p_feature_given_label(self, feature, label):
 		return self.featurecounts[label][feature] / self.labelcounts[label]
 
 	def p_feature(self,feature):
+		# This function is broken and wrong...
+		# It makes the assumption that the feature probabilities are class independent
+		# That's just not the case...
 		feature_total = 0.0
 		for label in self.labels:
 			feature_total += self.featurecounts[label][feature]