Navbot—College Navigation Chatbot Using Deep Neural Network

Abstract

AbstractIn terms of size, college campuses are massive. It is an uphill task to identify in which block a particular room is located or where the block is located within the campus. A chatbot has been developed to address this problem. A chatbot is an interactive artificial intelligence program that attempts to mimic human behavior by interpreting input and responding appropriately in text format. Deep learning model and natural language processing algorithms are used to develop the chatbot. The knowledge base for this chatbot will be in JavaScript Object Notation (JSON) format. Users can request college navigation-related queries. The query is processed using natural language processing techniques such as tokenization and lemmatization. The processed query after spell correction is given as an input to the Deep Neural Network (DNN) algorithm. The proposed chatbot will search the processed query in the knowledge base and respond with the corresponding answer using a sequential DNN model with five hidden layers. User interface of the chatbot is developed using Hyper Text Markup Language (HTML), Cascading Style Sheets (CSS) and Java Script. The proposed model will help in navigating people inside the college to different blocks through google map links and inside the blocks through textual directions. The model works with an accuracy of 98%.KeywordsDeep learningArtificial intelligenceNatural language processingChatbotNavigation