Naive Bayes

image

Naive Bayesian Learner

Signals

Inputs:

Data set

Preprocessed data

Outputs:

Naive Bayesian learning algorithm with settings as specified in the dialog. It can be fed into widgets for testing learners.

Trained classifier (a subtype of Classifier). The Naive Bayesian Classifier signal sends data only if the learning data (signal Data) is present.

Description

NaiveBayes Widget

The only option in this widget is the name under which it will appear in other widgets. The default name is 'Naive Bayes'. When you change it, you need to press 'Apply'.

Examples

Here we present two uses of this widget. First we compare the results of Naive Bayesian learner with another learner, a Random Forest.

image

The second schema show the quality of predictions made with Naive Bayes. We feed the Test Learners widget a Naive Bayes learner and then send the data to the Confusion Matrix. In this widget we select misclassified instances and show them in Scatterplot. The bold dots in the scatterplot are the misclassified instances from Naive Bayes.

image