Abstract

Experiencing problems without troubleshooting in the services that is offered by telecommunication operators causes the decreasing customer loyalty and loss of income. Data mining provides improved information through analysis of available data in the telecommunications industry. In this study, the data mining was applied on the troubleshooting process of broadband network in one of a leading telecommunication company in Turkey. In this scope, 4032 data that obtained from the company during the March-May 2019 period were used. 3748 samples were included the analysis after the pre-processing step. Seven different variables including Trouble Center, Work Order, Team Number, Service Type, Duration of Service, Complaint Type and Result data were recorded during the trouble recording process. According to the result, J48, PART and Multilayer Perceptron classifiers were performed better than others in the data set. The current research is important in terms of being a guiding work in ensuring effective control of processes in troubleshooting analysis.