Machine Learning for Classification of IT Support Tickets

The IT support teams are responsible for the correct functioning of IT resources, ensuring that people can work with the least amount of interruption or struggle. The first challenge of the help desk team is to associate a support request with its correct topic, minimizing the ticket resolution time and increasing user satisfaction. This paper presents a machine learning classification model to identify the topic of the incoming support ticket. Our model achieves 75% of average precision among 7 categories. The complete code, model, and anonymized data to reproduce this work were made available online. A web prototype (backend and frontend) was developed for IT analysts to use the system in their daily work.