A Multilingual iChatbot for Voice Based Conversation

Abstract

In the present pandemic crisis, chatbots are assisting consumers in resolving difficulties and consulting experts without the need for human intervention. Chatbot can serve them in analyzing their problem and delivering intelligent responses. In this paper, authors propose a multilingual iChatbot which allows users to communicate in six national languages and an international language. Along with text, it also supports questions through voice in their preferred language and receives responses in the same language, removing language as a barrier to communication. iChatbot can be utilized to answer relative questions of inquiring individuals in an interactive manner through voice command. A multilingual database utilizing bag of words algorithm and natural language processing techniques as a data processor is created. To make the chatbot intelligent feed forward deep neural network is used which enhances the overall performance of iChatbot for multilingual input. The proposed iChatbot outperforms the traditional method of chatbot and achieve accuracy more than 95%.