A STATISTICAL MODEL TO ENHANCE TOPIC DETECTION OF ARABIC TEXT

Emad Aloqayli, Yaseen Alquran, Dana Hazaimeh, Mahmood Qudah

Abstract: The exponential growth of the available Arabic documents online increased the need for techniques that help in classifying and processing these documents. The nature of the Arabic language and the presence of noisy information in the documents’ contents make it difficult to guarantee accurate results when performing different processes, like topic detection and classification. In this paper, a statistical model is proposed to enhance the ability of the topic detection of Arabic text. The model is evaluated using a document set. The results are measured using precision, recall, and accuracy and the preliminary indications show that the model is able to provide promising results.

Keywords: Arabic Language, Classification, Information Retrieval, Topic Detection

DOI: https://doi.org/10.15623/ijret.2017.0611003