A comparative study of Named Entity Recognition for Arabic using ensemble learning approaches

The ensemble learning has been successfully applied to many Natural Language Processing (NLP) tasks. For the Arabic Named Entity Recognition (NER) task, most studies in the literature have only focused on traditional classification methods and until now no one to the best of our knowledge has studied the ensemble learning for the Arabic NER task. In this paper, we apply six ensemble learning approaches to the Arabic NER task and we present a comparative study between these six ensemble learning approaches and six traditional classification approaches on two Arabic NER datasets (ANERcorp and AQMAR). The empirical results show that the ensemble learning methods significantly outperform the traditional classification methods. The Random Forests method achieves the best F1-measure results of 86.57% and 82.51% on each dataset, respectively.