NaiveBayes class -

Naive Bayes classifier

Description

A NaiveBayes object defines a Naive Bayes classifier. A Naive Bayes classifier assigns a new observation to the most probable class, assuming the features are conditionally independent given the class value.

Construction

NaiveBayes

Create NaiveBayes object

Methods

disp	Display `NaiveBayes` classifier object
display	Display `NaiveBayes` classifier object
fit	Create Naive Bayes classifier object by fitting training data
posterior	Compute posterior probability of each class for test data
predict	Predict class label for test data
subsasgn	Subscripted reference for `NaiveBayes` object
subsref	Subscripted reference for `NaiveBayes` object

Properties

CIsNonEmpty	Flag for non-empty classes
CLevels	Class levels
CNames	Class names
CPrior	Class priors
Dist	Distribution names
NClasses	Number of classes
NDims	Number of dimensions
Params	Parameter estimates

Copy Semantics

Value. To learn how this affects your use of the class, see Comparing Handle and Value Classes in the MATLAB Object-Oriented Programming documentation.

Examples

Predict the class label using the Naive Bayes classifier

load fisheriris

Use the default Gaussian distribution and a confusion matrix:

O1 = NaiveBayes.fit(meas,species);
C1 = O1.predict(meas);
cMat1 = confusionmat(species,C1)

This returns:

cMat1 =

    50     0     0
     0    47     3
     0     3    47

Use the Gaussian distribution for features 1 and 3 and use the kernel density estimation for features 2 and 4:

O2 = NaiveBayes.fit(meas,species,'dist',...
{'normal','kernel','normal','kernel'}); 
C2 = O2.predict(meas);
cMat2 = confusionmat(species,C2)

This returns:

cMat2 =

    50     0     0
     0    47     3
     0     3    47

References

[1] Mitchell, T. (1997) Machine Learning, McGraw Hill.

[2] Vangelis M., Ion A., and Geogios P. Spam Filtering with Naive Bayes - Which Naive Bayes? (2006) Third Conference on Email and Anti-Spam.

[3] George H. John and Pat Langley. Estimating continuous distributions in bayesian classifiers (1995) the Eleventh Conference on Uncertainty in Artificial Intelligence.