Chatbot for Admission Related Queries for Bahria University

Abstract

This project focuses on developing a chatbot for admission related queries for Bahria University. This chatbot can understand user queries in both English and Roman Urdu. The framework used to develop this chatbot is RASA. It works by extracting the features from the training data and then trained through machine learning algorithms and a model is created. The query is also processed similarly and classified to an intent along with entity extraction.A pipeline was configured which provided good results for both English and Roman Urdu . not only in the context of intent classification but also Entity extractionwith average accuracy, precision , recall and f1 scores were found to be 94 percent, 88 percent, 83 percent and 84 percent respectively. This document covers the process of collecting data and training and also the development of user interface and evaluation. This chatbot provides students with a single source for admission information.