Named entity recognition in Assamese: A hybrid approach

Most NER systems have been developed using one of two approaches: Rule-based or Machine-Learning, with their strengths and weaknesses. In this paper, we propose a hybrid NER approach which is a combination of both rule-based and ML approaches to improve the overall system performance for a resource poor language like Assamese. Our proposed hybrid approach is capable of recognizing four types of NEs: Person, Location, Organization and Miscellaneous. The empirical results obtained indicate that the hybrid approach outperforms both rule-based and ML when processed independently. The hybrid Assamese NER obtains an F-measure of 85%-90%.