An Adaptive Automated Arabic Essay Scoring Model Using the Semantic of Arabic WordNet

Abstract

An Automated Essay Grading (AEG) system is designed to be used in public schools, universities and companies, which is used to saving time and cost and to reduction in errors and unfairness due to human bias. The AEG widespread use and applied for multi-language such as English and French and others, but the researches in automated Arabic essay is limited. Therefore, this research introduces an Arabic Automated Essay Grading (AAEG). The aim of this research is to build a new model that is able to evaluate students' answers to Arabic essay questions with a score that is close to that provided by the teacher manually. The model relies on the semantic of Arabic WordNet model to achieve the accuracy and to avoid the weakness in stemming, after using a hybrid method in stemming with Arabic WordNet tables and using Arabic WordNet to search for all synonyms for all reference answer words so that students' marks don't oppress because they didn't write the same words reference answer.