Project Abstract

In 2017 it was found that 84% of companies have increased or started using artificial intelligence and Chatbot technology for customer service. The project uses this trend to try and bring an alternative to the traditional email system where the users have to wait for days to get the advice and guidance they need to resolve their queries. This project consists of making a chatbot for IT Services at Swansea University which helps users such as students and staff resolve their technical queries. To train the chatbot, A dataset relevant to IT Services at Swansea was generated, containing example queries & their variations along with their intent labels for the chatbot to identify the pattern for each query. The chatbot was implemented in Python with the help of the PyTorch & nltk libraries which helped with the advanced steps of natural language processing. This chatbot is then presented to the users via a UI where the user can then communicate with the chatbot. This chatbot model is then evaluated by a tool called the confusion matrix which helps visualise the precision and accuracy of the machine learning model. The confusion matrix portrays the model?��s classification accuracy for every intent in detail along with the scores of each class.

Keywords: Artificial Intelligence, Natural Language Processing, Machine Learning

 Conference Details

Session: Poster Session A at Poster Stand 132

Location: Sir Stanley Clarke Auditorium at Tuesday 7th 13:30 – 17:00

Markers: Julian Hough, Eike Neumann

Course: BSc Computer Science, 3rd Year

Future Plans: I’m looking for work