Naive Bayes Classifier

Wed, 24 Oct 2018 05:01:19 +0100

Introduction :

Naive Bayes is an extremely fast classification algorithm which uses Bayes Theorem as its basic building block. It assumes that the features or predictors in the dataset are independent.

The Bayes theorem is given as:

\[P(c|X) = \frac{P(c)P(X|c)}{P(X)}\]

where \(c\) denotes a class label and \(X\) is the predictor. The probabilities \(P(c)\) and \(P(X)\) are the prior probabilities of the class and the predictor. \(P(X|c)\) is the prior probability or likelihood of observing a feature \(X\) given class \(c\). \(P(c|X)\) is the posterior probability of the class \(c\) given a feature \(X\). Hence, the posterior probability of a class \(c\) givena a feature \(X\) can be found using different prior probabilities and likelihood which can be obtained from the existing dataset. In the plain english, the Bayes theorem can be stated as:

Classification on Amit Rajan

Naive Bayes Classifier

Introduction :