Language Identification Between Kannada And Malayalam Using Neural Networks

Abstract

               Language identification (LID) is one of the most widely used natural language processing technology. The language identification accurately identify the language from the given raw text corpus. It is one of the classification problem. Malayalam and Kannada Dravidian languages are used as raw text corpus which are morphologically rich in their features. In this work, we used deep learning network approach, focused on RNN, LSTM and GRU to improve the language modelling performance. We show that LSTM and GRU are achieve best performance compared to RNN. GRU is modified versions of RNN and it is similar to LSTM but fewer parameter. LSTM training time is too large than GRU. This paper introduces the language identification and mainly focuses deep neural network learning model for language identification.