Detection and Resolution of Anaphora in Myanmar Text using A Hybrid Approach

Abstract

Anaphora Resolution is the process of resolving references to an entity from the set of referring expressions or discourse. Anaphora resolution plays an important role in the natural language understanding. It is needed to be addressed in almost every application dealing with natural language such as language understanding and processing, dialogue interpretations, machine translation, question-answering, information extraction, text summarization, and so on. Anaphora resolution in English and other languages has been well done in early. But Myanmar language has not sufficiently applied. The paper presents a hybrid approach to resolve pronominal anaphora in Myanmar text. The hybrid model adopted the strategy based on the combination of a rule-based and machine learning approach. Rule-based manner with morphological information is used to collect anaphora and possible antecedents. K-Nearest Neighbor (k-NN) approach is used to select the most probable candidate as the antecedent of the anaphor.