Q-Meter: Quality Monitoring System for Telecommunication Services Based on Sentiment Analysis Using Deep Learning

Area of Science:

• Telecommunication Network Monitoring
• Quality of Experience (QoE) Analysis
• Social Media Data Analytics

Background:

• Network operators need effective Quality of Experience (QoE) monitoring systems.
• Subscriber complaints are a key indicator of service quality issues.
• Online Social Networks (OSNs) offer a rich source of user feedback.

Purpose of the Study:

• To propose Q-Meter, a novel system for enhanced subscriber complaint detection in telecommunication services.
• To leverage Online Social Networks (OSNs) and deep learning for real-time quality monitoring.
• To identify and analyze geographical areas with poor signal strength and coverage.

Main Methods:

• Utilizing deep learning, specifically Convolutional Neural Network (CNN) and Bidirectional Long Short-Term Memory (BLSTM)-Recurrent Neural Network (RNN) with soft-root-sign (SRS) activation, for sentiment analysis of OSN posts.
• Extracting subscriber geographical locations from OSN data.
• Analyzing complaint regions using a freeware application with Radio Base Station (RBS) information from open databases.

Main Results:

• Achieved 97% precision in classifying weak signal topics using the developed deep learning model.
• Determined that 78.3% of complaints are related to weak coverage.
• Validated coverage problems in 92% of identified complaint regions for a specific cellular operator.

Conclusions:

• Q-Meter offers a cost-effective and easily integrable solution for telecommunication quality monitoring.
• The system effectively detects subscriber complaints related to signal strength and coverage issues using OSN data.
• Q-Meter enhances network monitoring capabilities for current and future cellular networks, improving user QoE.