Authors : Hamzah Noori Feje and Mohanaed Ajmi Falih
Abstract: Because of the growing number of electronic documents, human being are badly in need of more rapid techniques for evaluating the link of documents. Summarization is representation of underlying written text. A full underst anding of the document is essential to form an ideal summary. However, achieving full underst anding is either difficult or impossible for computers. Therefore, selecting main sentences from the original text and introducing these sentences as a summary present the most frequent techniques in automated text summarization. This study propose using key phrase extraction module is applied to extract main important key phrases from the text that helps specify the most important sentences and find similar sentences based on similarity algorithm. It is applicable to extract one sentence from a set of similar sentences while overcoming the other similar sentences (i.e., sentences that have a greater similarity than the predefined threshold). This model is designed for single-document Arabic text summarization. The Recall-Oriented Understudy for Gisting Evaluation (ROUGE) matrix is employed for the assessment. For the summarization dataset, Essex Arabic Summaries Corpus was used. It has many topic based articles with multiple human summaries. This model achieved accuracy more than 80%.
Hamzah Noori Feje and Mohanaed Ajmi Falih, 2018. Automatic Summarization Arabic Text Using Key Phrases Extraction. Journal of Engineering and Applied Sciences, 13: 1395-1399.