Thesis (M.Sc.) - Cairo University - Faculty of Computers and Information- Department of Information Technology

In the last few decades, there was a dramatic increase of the content over the Internet, in- crease in Internet users, and increase demand for information that traditional information processing methods can not satisfy. With this boom of the Web{u2019}s content, an inevitable need for e{uFB00}ective techniques to exploit and make use of this huge amount of information is raised. In particular, the possibility of extracting the most important points from large piece of text is called text summarization. Text summarization (TS) task is still an active area of research in natural language processing. Several methods and techniques that have been proposed in the literature in di{uFB00}erence languages to solve this task have presented mixed success. Summarization process can be seen as a two-steps process: identifying the important parts of the whole text, and then generate the summary from these parts. In general, an automatic text summarization can be divided into two categories: extractive and abstractive text summarization. Furthermore, both extractive and ab- stractive text summarization can be used to summarize either single document or multi- document. However, such methods developed in a multi-document Arabic text summa-rization are based on extractive summary and none of them is oriented to abstractive summary. This is due to that abstractive techniques require a deeper language process that could be involved sentence reduction, information merging and language generation. In addition, the characteristic of Arabic language and lack of resources and preprocessing tools still challenging problem. In this thesis, a minimal language-dependent processing abstractive Arabic multi-document summarizer has been presented. The proposed model is based on textual graph to remove multi-document redundancy and generate coherent summary

Issued also as CD