Collaborators: Ludovico Francia
Starting from over 28,000 support tickets received by an IT department of a company, our research employs NLP techniques to automate the assignment of each ticket to a specific topic (thereby streamlining the routing to the appropriate team) and to predict the priority of each ticket by comparing the performance of various models.
Dataset: https://www.kaggle.com/datasets/tobiasbueck/multilingual-customer-support-tickets/data
- Final Report (PDF): A comprehensive write-up of our methodology, experiments, results, and conclusions.
- Jupyter Notebook: All the code used for text preprocessing, topic modeling, and priority classification.