Support ticket categorization through Latent Dirichlet Allocation

In contemporary corporate environments, the efficient handling of service desk tickets is pivotal for ensuring smooth IT operations. Support tickets, encompassing a range of customer requests and issues, are essential communication tools between customers and support teams. Proper categorization and swift resolution of these tickets are crucial tasks, ensuring customer satisfaction, productivity, compliance with service level agreements, and cost efficiency. In this study, the potential of Latent Dirichlet Allocation (LDA), an unsupervised machine learning algorithm, was explored for automatic categorization of service desk queries. Leveraging firsthand experience as an IT intern at Lavazza and access to a substantial dataset of tickets, this research successfully implemented LDA using Dariah-DE Topics Explorer. The findings indicate that LDA holds promise in enhancing ticket resolution efficiency, potentially revolutionizing service desk operations. However, the system's effectiveness depends on the quality and diversity of the training dataset and requires continuous optimization. Future research could focus on integrating LDA within existing service desk software, augmenting their capabilities for streamlined and efficient ticket handling processes. This study lays the foundation for further advancements in automating service desk ticket categorization, aiming for more advanced and responsive service desk operations in the future.