A Semantic Text Summarization Model for Arabic Topic-Oriented

Abstract

In the era of data overloading, Text Summarization systems (TSs) is one of the important Natural Language processing applications. These systems provide a concise form for the input document(s). According to the type of output summary, Text Summarization can be classified into extractive and abstractive. While the extractive text summarization is the process of identifying the important sections of the input text and producing them verbatim, the abstractive text summarization produces a new material in a generalized form. To facilitate the topic-oriented summarization, current research efforts focus on query-based text summarization, which summarizes the input document according to the user query. Although, the Arabic language is one of the Semitic languages and is spoken by 422 million people, there are very limited research efforts in Arabic query-based text summarization. In this paper, we propose a new Arabic query-based text summarization model. The model accepts both user query and Arabic document and then generates the extractive summary. The proposed model generates the extractive summary for the input document semantically by applying the Latent Semantic Analysis technique and exploiting the Arabic WordNet (AWN) ontology. Finally, to show the importance of the proposed model, a case study is presented.