OVERVIEW AND COMPARISON OF THREE CLASSIFIERS: ARABIC DOCUMENTS AS A CASE STUDY

Abstract

Nowadays, text classification is used in various fields of research and applications, such as information retrieval, text mining, and data mining. This study tests the Naïve Bayes, K-Nearest Neighbors, and Support Vector Machine algorithms on a relatively large dataset of Arabic documents. This dataset comprise 1,000 Arabic documents that are distributed across 10 classes. This comparison is based on recall and precision measures. The evaluation results show that the Support Vector Machine algorithms classifier outperforms the other two