Predicting Ticket Resolution Time in IT Support Systems: A Machine Learning Approach

The evolution of support ticket management in software companies has become crucial in the era of digitization, where the volume of reports is constantly increasing. Automating ticketing systems has become imperative to improve technical support management and meet growing customer needs. This study addresses two key questions: first, whether by analyzing historical information from a ticketing system, it is possible to predict resolution times in advance, and how features influence such predictions. Second, how predictive algorithms can be integrated into operational processes to proactively manage spikes in activity, optimizing resource allocation and improving workload management. Using data provided by CINECA, this paper aims to identify key factors in predicting the time to resolution of a ticket, leveraging both existing and temporal features specifically introduced for this analysis. The goal is to provide useful insights to improve efficiency and customer satisfaction in the area of service ticket management.