Abstract
Increasingly interested in research communities, the text classification area enables the text or part of the text to be classified into classes for extracting useful information. Expensive to scale, the manual classification tasks are becoming vulnerable to potential unreliability as documents in the world increase, especially if the classes number more than two (multiclass classification). As a classification technique based on algorithms, automatic classification facilitates the automatic categorization of text documents to classes, thus resulting in reliable and efficient classification. This paper aims to describe the process of using the Naive Bayes classifier for text classification with one-of and multiclass, especially in cases where the probability of imbalanced classes is higher. Our proposed process consists of a number of steps such as data preprocessing, classification model building, evaluating and predicting classes as final classification results.