Usage of Naive Bayes

Text classification is one of the strongest characteristics of the Naive Bayes classifier. Additionally, Naive Bayes has also other advantages, when you use it for domains.

The other advantages of the Naive Bayes classifier are as follows:

You can quickly create classification models.
You can frequently create many classification models.
This kind of creation might be necessary, for example, because the classification model is quickly outdated, or because new classification tasks appear dynamically.
You have minor computational effort.
You do not have to tune the parameters.
You do not have additional effort regarding over-fitting prevention.

In some areas, however, other algorithms are more refined in efficiency and in accuracy.

Examples for these areas are:

Classification tasks with many attributes, each of which might have some marginal impact on the class
Classification tasks that do not have strong relationship patterns that you can use for prediction
The prediction is based on the values of few of the most influential attributes.