Naive Bayes - Distributions Tab

Select the Distributions tab of the Naive Bayes dialog box to access options to specify the conditional distributions of the independent variables that are used to calculate the conditional posterior probabilities for making predictions. These options become available after variable selection has been made.

Threshold

When the frequency of a categorical level of an independent variable is zero with respect to one of the categorical levels of a dependent variable, the likelihood of the Naive Bayes model will always yield zero, which is rather superficial. In the Threshold field, you can specify a threshold parameter that specifies a lower bound to use in stead when such frequencies are encountered by the Naive Bayes model.

Select distributions

Click this button to display a standard variable selection dialog box, which is used to specify the type of the distributions you want to assign to the continuous independent variables. These include Normal, Lognormal, Gamma, and Poisson. Variables designated to have a Poisson distribution must be selected as a continuous (ordinal) variable first. Note that categorical variables can only have discrete probabilities.

Distribution list

The list at the bottom of the dialog box displays the independent variables and distribution types.