Deep Learning Models Applied to Prediction of 5G Technology Adoption

Abstract

The issue addressed by this research study is the public’s scepticism about the benefits of adopting 5G technology. Some have even gone so far as to say that the technology can be harmful to people, while others are still looking for reassurance. This is why it is crucial to comprehend the primary factors that will affect the spread of 5G networks. The method used for this heavily relies on a deep learning algorithm. Channel metrics, context metrics, cell metrics, and throughput data are the conceptualized variables that will serve as the primary indicators for determining the adoption of 5G technology. Three deep learning models—deep reinforcement (DR), long-short term memory (LSTM), and a convolutional neural network (CNN)—were applied. The results show that the DR model and the CNN model are the most effective at predicting the elements that would affect 5G adoption. Despite the fact that LSTM models appear to have a high degree of accuracy, the quality of the data they output is quite poor. However, this is the case even when the models appear to be rather accurate. The logical inferences drawn from these findings show that the DR model and the CNN model’s applicability to the problem of predicting the rate at which 5G will be adopted can be put into practice with a high degree of accuracy. The novelty of this study is in its emphasis on using channel metrics, context metrics, cell metrics, and throughput data to focus on predictions for the development of 5G networks themselves and on the generation of the elements that determine the adoption of 5G. Previous efforts in the literature failed to establish methods for adopting 5G technology related to the criteria considered in this study; hence, this research fills a gap.

Keywords

• 5G technology
• 5G adoptions
• deep reinforcement
• long short-term memory