Efficient Relay Selection in NOMA-Based Cooperative Networks: A Supervised Learning Approach With Random Forest

In this research, we examine how machine learning is used to cooperative communication networks' relay selection, emphasizing transmission under non-orthogonal multiple access. We introduce a novel relay selection algorithm based on Random Forest and compare its performance with SVM method previously discussed in this context and enhanced two- stage max-min selection rule. According to the study results, the Random Forest algorithm shows improved outage probability and system throughput even when the signal regime is more difficult. The superiority of the presented algorithm lies in its ability to combine NOMA and OMA transmission sources in an intelligent way, as well as ensure adaptability even in more complex situations due to high-dimensional classification. This work provides readers with valuable experience in applying machine learning methods to improve communication systems.