Design and Evaluation of a Machine Learning-Based Model for Automated Incident Classification in IT Helpdesk Systems

Abstract

Automated incident classification in IT helpdesk environments holds the potential to significantly enhance response efficiency and reduce human error. This paper presents the design and evaluation of a machine learning-based system tailored for classifying helpdesk tickets by incident category. Leveraging historical ticket data from enterprise IT support logs, several models, including Random Forest, Support Vector Machines, and Multinomial Naïve Bayes, were trained and benchmarked. Results demonstrate that the Multinomial Naïve Bayes model achieved the best performance, with an overall classification accuracy of 84.3%. This study contributes to the growing literature on applying supervised learning techniques to IT Service Management (ITSM) and supports the use of lightweight models for real-time ticket triaging.