Arabic text summarization based on graph theory

Automatic text summarization is a process of reducing the length of original document without affecting the content by extracting important information from huge amount of text data. The main goal is to facilitate the task of reading and searching information in large documents. Text summarization is the most challenging task in information retrieval especially for Arabic language. Unlike English and European languages, researches in Arabic text summarization are very few and still in their beginning. Summarization systems for Arabic are however still not as mature and as reliable as those developed for languages like English. In this paper, we introduce a new method for Arabic text summarization based on graph theory and semantic similarity between sentences to calculate importance of each sentence in document and most important sentences are extracted to generate document summary. In addition, because words sharing a root are semantically related, feature selection techniques based on the root can improves the semantic similarity between sentences and increases the weight of the semantic feature in the sentence. We will first review the related works in this field and especially in Arabic text summarization. Then we will present the architecture of our system, its components and its features. The last section will evaluate the system and compare it to other existing methods.