Leveraging Unsupervised Machine Learning for New Traffic Types Classification in Hybrid Satellite Terrestrial Network
Résumé
By offering real-time insights, automating network
modifications, and improving network performance,
unsupervised machine learning can be a useful tool
for managing novel traffic types in hybrid terrestrialsatellite networks. It aids network operators in adjusting to shifting traffic patterns and addressing the
needs of a dynamic and ever-changing communication landscape. We investigate the possibility of unsupervised learning based on four different learning
algorithms created for identifying novel traffic with
the aim of establishing a novel framework that maximizes proactive resource allocation. Based on a
performance assessment that was done, the findings
show that that the K-Means Clustering and Gaussian Mixture Model provide the best report between
acceptable silhouette and Principal Component Analysis.