Deep Learning based Campus Assistive Chatbot

With the rapidly increasing dependence and expectations on online platforms to provide faster responses during all hours of the day, a web-based chatbot could be considered as the ideal solution to do so in the educational domain. A chatbot is a conversational agent capable of emulating human interactions. The purpose of this project is to implement a chatbot for a college website in order to make it more accessible, convenient and to enhance user experience. Campus navigation through Google Maps is an additional feature which is integrated into the chatbot to better assist the user. A menu containing frequently used links is provided which redirects the webpage according to the user’s choice. Using Artificial Intelligence and Deep Learning technologies, a chatbot EVA was created. Natural Language Processing is incorporated to allow the user to communicate with the chatbot using natural language input. A Sequential Model is developed using Neural Networks and the concept of Bag of Words is utilized. Tkinter is the tool used to set up the Graphical User Interface. A basic level performance test was conducted which revealed an efficiency rate of 89%. Finally, the chatbot was integrated with a college website which resulted in the generation of a functional college chatbot. EVA not only provides concise answers to the user’s queries, but also necessary links to information regarding course content, extracurricular activities as well as directions to various locations on campus as per the user’s requirement.