Just A Rather Very Intelligent Chatbot
Software Engineering Project
Abstract
Natural Language Processing has found wide applications from language translation to text generation. Automated chatbots are one such major application that has made its way through everyday lives. Most of the conversational agents do not take the user's emotions into account. Whats even worse is that they provide a list of possible queries that the user must adhere to. We make an attempt to develop an android conversational agent that is more personalized for the user. The agent uses a Bayes classifier to detect the emotional state of the user and based on the predicted emotion, the appropriate conversational agent i.e. text generator is used. We use simple seq2seq models for text generation. The android client sets up a reliable TCP channel with the python server and the user authentication and personal details, and conversation history are stored in a relational database.
Report & Code
