A Data-Driven Framework for AI-Powered LTE Network Performance Optimization

Abstract

Optimizing modern LTE networks in multi-vendor environments such as Djezzy’s presents significant challenges. This project focuses on integrating AI-driven KPI forecasting and prediction capabilities. A scalable data warehouse with a dimensional model supports efficient KPI data management through robust ETL processes. Advanced AI models, including LSTM and Prophet, were implemented for time series forecasting to enable proactive prediction of network performance metrics. These predictions are delivered via a dedicated web page to enhance network planning and decision-making. Evaluation on real-world data demonstrates improved foresight for network operations, marking a step forward toward data-driven LTE management in complex multi-vendor contexts