Please enable JavaScript.
Coggle requires JavaScript to display documents.
Naive Bayes - Coggle Diagram
Naive Bayes
Multinominal
Algorithm
- Calculate Prior Probabilities for each class;
-
-
- Calculate likelihood for each value in each predictor variable for each class
-
-
- Calculate P(classi|test data) ∞ P(test data value1|classi) x P(test data valuen|classi) x P(prior for classi) ; where n=num of variables in test data
- Compare P(classi|test data) for each class and take the class with largest likelihood for the test data
-
Gaussian
Algorithm
- Find Gaussian distributions for each variable in dataset
-
-
Likelihood=P(predictorj |classi)=y-axis coordinate on predictor variablej Gaussian curve of classi that corresponds to predictorj value [x-axis coordinate]
-
- Calculate P(classi|test data) ∞ P(test data value1| classi) x P(test data valuen|classi) x P(prior for classi) ; where n=num of variables in test data
-
-
- Calculate Prior Probabilities for each class;
-
-
- Compare P(classi|test data) for each class and take the class with largest likelihood for the test data
-
-
-
