A Connectionist Approach to Learn Marathi Language

In this paper, we have investigated the inductive inference of complex grammar of subset of Marathi (Indian) language and results are reported. We have investigated Elman recurrent networks (ERNs), Jordon recurrent networks (JRNs), time lagged recurrent networks (TLRNs) and recurrent neural networks (RNNs). In this empirical study, we consider the task of classifying Marathi language sentences as grammatical or ungrammatical as well as modeled the problem as a prediction problem. We have also analyzed the operation of the networks by investing rule approximation.