Channel Selection Strategy in 5g Based on Contextual Mab

5th Generation (5G) has become an essential part of humans’ lives, much research has been done in network areas like 2G, Long-Term Evolution (LTE) channel selection, but less research has been done on 5G channel selection; hence, this study aims to provide an insight into the 5G channel selection issue. This study uses a contextual Multi-armed Bandit (MAB) algorithm to select the best arm based on contextual features, and the main contribution is constructing a methodology to select a 5G channel in a real situation to improve people’s quality of life. The conclusion drawn from the experiment is that the best arm can be selected well through context features, but due to the number of channels in the dataset is not sufficient and the algorithm strategy and parameter settings is not optimal, the cumulative regret shows a linear trend rather than a sublinear trend and eventually reaches about 3000. However, the selection strategy used in the algorithm is very effective, causing the optimal arm selection rate to rise sharply in the early stage and quickly reach about 0.99.