Precise 5G Traffic Forecasting by Using Search-Economics Algorithm to Fine-Tune the GRU Weights
摘要
While mobile communication technology has seen significant advancements, it struggles with overloads during high traffic periods, leading to rapid deterioration and necessitating hardware upgrades. To address this, traffic forecasting is crucial. Existing methods rely on regression models, often yielding suboptimal results due to the complex nature of network traffic. Recurrent neural networks-based method has improved this, but overfitting remains a concern. This paper aims to enhance deep learning predictive accuracy using an innovative heuristic algorithm to optimize network weights.