Pronominal Anaphora Resolution in Hindi Language Using Number Agreement and Animistic Knowledge

Anaphora resolution is always been one of the most active research areas in the domain of Natural Language Processing (NLP). Anaphora resolution refers to the act of referring, it is commonly used for pronoun resolution in discourse. It tends to figure out who or what the referent is used for. Much of the work is been carried out and still going on in this field. A number of algorithms has been proposed to solve and to find out the referent. Anaphora resolution approaches rely on domain knowledge and linguistic. The disadvantage of such approach is that it takes lot of effort and hence is laborious. This paper focuses only on pronominal anaphora resolution for Hindi language sentences that are preprocessed by a part-of-speech tagger. Inputs are checked using number agreement and animistic of the noun phrase. The paper defines resolving anaphora on the basis of limited knowledge of the language. A method called transitive anaphora resolution is also proposed, where the antecedent may not be present in the same or the previous sentence.