Energy and Spectral Efficiency Optimization with Security in Massive MIMO–5G Using Integrated Machine Learning

With the rapid expansion of 5G networks, massive MIMO has become essential for achieving higher data rates and increased capacity.However, maintaining an optimal balance among energy efficiency (EE), spectral efficiency (SE), and network security in dense multicell environments remains a significant challenge due to dynamic interference patterns, complex power allocation requirements, high-dimensional channel data, and the increased vulnerability of dense networks to cyber threats and anomalous behaviour.Additionally, conventional optimization and detection methods often fail to scale or adapt effectively in real-time scenarios.This study aims to develop an AI-driven framework capable of simultaneously optimizing EE and SE along with strengthening security in massive MIMO-5G systems.The proposed framework integrates intelligent power allocation, adaptive resource management, and AI-based anomaly detection.Hybrid 5G datasets are analyzed using Deep Neural Networks (DNNs), Random Forest models, and multi-objective optimization techniques.Advanced learning models are further incorporated to enhance security and system reliability.The DNN classifier achieved 100% accuracy in distinguishing between secure and insecure transmissions, whereas the Random Forest model achieved a predictive accuracy of R² = 0.9946.The AI-assisted power allocation scheme improved EE by 35% and SE by 28% compared to traditional approaches.The anomaly detection module achieved F1-scores above 0.92.Additional system gains included a 40% reduction in latency, a 45% increase in throughput, and improved resource utilization across multi-cell massive MIMO deployments.The framework provides a unified, data driven solution to enhance EE, SE, and network security.It supports scalable deployment, improves real-time decisionmaking, and offers significant performance gains over conventional massive MIMO optimization methods.