Automatic Summarization of Arabic Text Using Ontology
Keywords:
Automatic Sumarization, Arabic text summarization, Extractive Summarization, Abstractive Summarization, Semantic Summarization, Arabic Wordnet, Ontology, Arabic Encyclopedia, MappinAbstract
Text summarization is the process of extracting or compiling important information from the original texts and presenting them in a summary form. Summarizing helps in obtaining the required information in the least time.
There has been a lot of research, studies and systems for the automatic summarization of texts in various European and other international languages, but research and studies in the field of summarizing Arabic texts are few and growing slowly (El-Haj et al., 2011) Most of them followed the extractive approaches and the use of statistical techniques and moved away from the abstract approaches and the use of semantic techniques based on knowledge sources such as ontology. Therefore, in this paper, we proposed an automatic summarization system for Arabic texts by adopting ontology, which is a general, monolingual and single-document extractive summarization system that uses semantic techniques based on ontology. Semantic relations between concepts derived from the Arabic ontology, which are synsets, holo_part, hyponym, and holo_member. The widely used Arabic Wordnet ontology has been used after expanding and enriching it with concepts and relationships derived from the Arabic encyclopedia using the mapping technology.The system was evaluated using Essex Arabic Abstracts Collection (EASC) (El-Haj et al.,2010), and measures of Precision, Recall and F measure. The results showed that the average precision, recall, and F measure in summarizing using the proposed ontology are good proportions compared to the proportions of automated summarization systems for Arabic texts.