Rule based anaphora resolution in Hindi

Handling of human languages, especially written texts, requires analysis and implementation at different linguistic levels such as morphological analysis, Parts-of-Speech(POS) labeling at word level; chunking and clause identification at word group level; syntactic parsing at the sentence or structural level; semantic analysis at the level of meaning and finally discourse analysis at the discourse or text level. Discourse level analysis involves various sub-problems that deal with relationships between sentences and larger linguistic units. One such phenomenon is Anaphora Resolution (AR). Anaphora occurs frequently in written texts and spoken languages. Anaphora resolution is required in almost every application of NLP as information extraction (IE), summarization and Machine translation. The main focus of this paper is Entity Resolution (ER) and pronominal forms. Various approaches have been discussed to resolve Entity-pronoun references in Hindi. This paper will discuss the Rule Based approach among all the approaches, used to identify the anaphors and their antecedents. Rules have been framed for five types of pronouns that have been discussed in this paper.