<![CDATA[This study focuses on developing a ticket classification model for the Service Desk at an insurance company to enhance operational efficiency. Manual ticket classification is time-consuming and prone to errors, so the research aims to compare the performance of various classification algorithms to determine the best model. The methodology involves text mining and machine learning techniques using four main algorithms: Random Forest, Decision Tree, Support Vector Machine (SVM), and Naïve Bayes. The data comes from Service Desk tickets processed through text preprocessing stages. Findings indicate that the Random Forest model with a combination of TF-IDF Unigram features in the Access context performs best in classifying IT Support tickets, with a Precision of 0.76%, Recall of 0.66%, F-Score of 0.70%, and Accuracy of 0.54%. Implementing this model is expected to improve operational efficiency and user satisfaction with IT services, speeding up ticket handling, reducing administrative workload, and enhancing user satisfaction with IT services.]]>
Copyrights © 2024
