Optimizing 5G Congestion Control with Machine Learning-Based Predictive Models

As 5G networks continue to grow, ensuring smooth and efficient data transmission has become more important than ever. With increasing network traffic, advanced congestion control mechanisms are needed to maintain reliability and performance. This paper explores how machine learning can help predict and manage congestion in 5 G environments using a combination of supervised and unsupervised learning techniques. Our approach is built on a diverse dataset that includes key network factors such as congestion window size, throughput, queue size, and packet loss, capturing the ever-changing nature of 5 G systems. To improve accuracy, we apply data preprocessing and feature engineering, which are especially crucial for unsupervised learning. Unsupervised techniques help uncover hidden patterns in network behavior, while supervised algorithms refine these insights to make precise congestion predictions. By combining these methods, our framework provides a smarter and more adaptive way to handle congestion in 5 G networks. This research highlights how machine learning can address real-world network challenges, paving the way for more intelligent and efficient communication systems. In our study, we tested 20 supervised and five unsupervised algorithms, ultimately identifying the most effective based on key performance metrics.