Content area
The complexity of wireless network service issues, combined with the limited capabilities of troubleshooting tools and varying troubleshooting knowledge among telecom care agents, has led to a situation where care agents frequently perform repetitive subscriber profile updates on the core network during troubleshooting. This has resulted in a spike in core network capacity utilization. As the number of subscribers has increased and wireless industry applications have evolved beyond voice calls, care agents have increasingly relied on trial-and-error troubleshooting methods to meet ticket resolution service level agreement and minimize service downtime. Unfortunately, this approach has generated unnecessary traffic to the core network, leading to inefficient network capacity utilization.
This praxis developed a predictive model to indicate the solution parameter for resolving the issue by analyzing the minimal network capacity utilization. The model was built by studying the nature of issues subscribers reported, the suggested steps, and the network capacity impact for that fix flow. Based on these inputs, a supervised machine-learning model was developed, trained, validated, and tested. This model will enable care agents to apply the fixed parameter for the issue that not only solves the subscriber issue but ensures the minimum network capacity utilization by eliminating the redundant operation that results network utilization spike.