This article proposes a novel meta-learning-based hyperparameter optimization framework for wireless network traffic prediction (NTP) models. The primary objective is to accumulate and leverage the acquired hyperparameter optimization experience, enabling the rapid determination of optimal hyperparameters for new tasks. In this article, an attention-based deep neural network (ADNN) is employed as the base-learner to address specific NTP tasks. The meta-learner is an innovative framework that integrates meta-learning with the k-nearest neighbor algorithm (KNN), genetic algorithm (GA), and gated residual network (GRN). Specifically, KNN is utilized to identify a set of candidate hyperparameter selection strategies for a new task, which then serves as the initial population for GA, while a GRN-based chromosome screening module accelerates the validation of offspring chromosomes, ultimately determining the optimal hyperparameters. Experimental results demonstrate that, compared to traditional methods, such as Bayesian optimization (BO), GA, and particle swarm optimization (PSO), the proposed framework determines optimal hyperparameters more rapidly, significantly reduces optimization time, and enhances the performance of the base-learner. It achieves an optimal balance between optimization efficiency and prediction accuracy.
This paper proposes an interpretable user-behavior-based (UBB) network traffic prediction (NTP) method. Based on user behavior, a weekly traffic demand profile can be naturally sorted into three categories, i.e., weekday, Saturday, and Sunday. For each category, the traffic pattern is divided into three components which are mainly generated in three time periods, i.e., morning, afternoon, and evening. Each component is modeled as a normal-distributed signal. Numerical results indicate the UBB NTP method matches the practical wireless traffic demand very well. Compared with existing methods, the proposed UBB NTP method improves the computational efficiency and increases the predictive accuracy.
1 January 2026 in IEEE Transactions on Vehicular Technology