Automatic Essay Scoring for Arabic Short Answer Questions using Text Mining Techniques

Automated Essay Scoring (AES) systems involve using a specially designed computing program to mark students’ essays. It is form of online assessment supported by natural language processing (NLP). These seek exploit advanced technologies reduce the time and effort spent on exam scoring process. have been applied in several languages, including Arabic. Nevertheless, applicable NLP techniques Arabic AES are still limited, further investigation needed make suitable for achieve human-like accuracy. Therefore, this comparative empirical experimental study tested two word-embedding deep learning approaches, namely BERT Word2vec, along with knowledge-based similarity approach; WordNet. The used Cosine measure provide optimal student answer scores. Several experiments were conducted each proposed approaches available short question datasets explore effect stemming level. quantitative results indicated that models contextual embedding can improve efficiency as meaning words differ different contexts. serve catalyst future research based models, approach achieved best Pearson Correlation (.84) RMSE (1.003). However, area needs increase accuracy become practical system.