Most Named Entity Recognition (NER) systems follow either a rule-based approach or machine learning approach. In this paper, we introduce out attempt at developing a hybrid NER system, which combines the rule-based approach with a machine learning approach in order to obtain the advantages of both approaches and
overcomes their problems. The system is able to recognize eight types of named entities including Location,
Person, Organization, Date, Time, Price, Measurement and Percent. Experimental results on ANERcorp dataset indicated that our hybrid approach outperforms the rule-based approach and the machine learning approach when
they are processed separately. Moreover, our hybrid approach outperforms the state-of-the-art of Arabic NER.
