An Ontology-based Summarization System for Arabic Documents (OSSAD)

摘要: With the problem of increased web resources and huge amount information available, necessity having automatic summarization systems appeared. Since is needed most in process searching for on web, where user aims at a certain domain interest according to his query, domain-based summaries would serve best. Despite existence plenty research work English, there lack them Arabic due shortage existing knowledge bases. In this paper an Ontology-based Summarization System Documents, OSSAD, introduced. Domain extracted from corpus represented by topic related concepts/keywords lexical relations among them. The user’s query first expanded using WordNet then adding domain-specific base expansion. For summarization, decision tree algorithm (C4.5) used, which was trained set features original documents. testing dataset, Essex Summaries Corpus (EASC) used. Recall Oriented Understudy Gisting Evaluation (ROUGE) used compare OSSAD with human along other systems, showing that proposed approach demonstrated promising results.