Question

Discuss the main differences between Naïve Bayes Classifier and Softmax Classifier.

Assess when will you use Naïve Bayes over Softmax Classifier

Please provide at least 3 differences thx

Answer 1

Before differentiating the two classification i want to clear the definition of class-ification which means that based on certain parameter of a object we put then in one or more class.

1. Naive Bayes Classifier, has a assumption which isn't there in the Soft max Classifier which means, the assumption is the reason why it is supposed to be "naive". Assumption is - Naive Bayes Classifier assumes that one parameter didn't had anything to do with the other parameter. SoftMax doesn't have any such assumption.

2. Softmax Classifier, has a implementation based on the activation and not a direct Bayes theorem application like Bayes Classifier. The activation and the normalisation of the Z(l) in the SoftMax Classifier is the unique part neither of those are present in the Navie Bayes Classifier.

3) The structure of the SoftMax Classifier is implemented with the neural network of the same, and the output layers represent the normalized probability of the possibilities. Where as the naive base theorem will give a option which is taken case separately and don't need to have a structural network but will work on the just the simple formulaptic regression.

4) There is no such boundary of naive base theorem result. the output of the SoftMax Classifier is actually a linear boundary point.