User Experience Estimation in Multi-Service Scenario of Cellular Network

Area of Science:

• Computer Science
• Telecommunications Engineering
• Network Management

Background:

• User experience estimation in wireless networks is crucial for network automation.
• Current methods lack in-depth analysis of user experience decline factors.

Purpose of the Study:

• To propose an optimized naive Bayesian classifier for accurate user experience estimation.
• To integrate user experience prediction with network fault diagnosis.
• To develop a cell-level user experience evaluation standard.

Main Methods:

• A two-step optimization for kernel function and bandwidth in a naive Bayesian classifier.
• Categorizing Key Performance Indicator (KPI) data into five groups for estimation.
• Employing a voting mechanism for final estimation and feedback on degrading KPIs.
• Summarizing user experience for three main services for cell-level evaluation.

Main Results:

• The optimized method significantly improves the accuracy of user experience estimation.
• The voting mechanism effectively identifies KPIs causing user experience degradation.
• The approach enables timely diagnosis of abnormal network values.
• Accurate cell-level user experience evaluation was achieved.

Conclusions:

• The proposed integrated approach enhances wireless network management efficiency.
• Accurate user experience estimation and fault diagnosis are achievable.
• The method provides valuable insights into user experience influencing factors.