Исследование методов прогнозирования трафика в сетях 5G на основе глубокого обучения

2 с.

This study provides a robust solution based on artificial intelligence (AI) methods to overcome problems in 5G network traffic that can affect overall performance. Predicting network traffic using artificial intelligence techniques such as Deep Learning. The rapid growth of the wireless network leads to loss of control in the network and the network goes beyond the originally projected parameters, including the maximum time error detection. Network traffic prediction allows 5G networks for proactive response to events before they happen. With the fast increase in the amount, quality, and detail of traffic data, so, there is a need for new techniques that can use the information in the data to provide better results while they have the ability to scale and cope with increasing amounts of data. The desire of telecom operators and service providers to provide new services to customers. Deep learning allows you to predict the loading of network resources and apportion them accordingly. Also, the proposed deep learning can improve the prediction better than traditional time series, besides it more efficient and faster to fit data compared to traditional time series models, in addition, they deal with larger data sets better than traditional time series models.

5g, artificial intelligence, deep learning, prediction

Русский

75-я Научно-техническая конференция Санкт-Петербургского НТО РЭС им. А.С. Попова, посвященная Дню радио, Санкт-Петербург, 20–24 апреля 2020 г. : труды конференции / Санкт-Петербургская организация Общероссийской общественной организации «Российское научно-техническое общество радиотехники, электроники и связи им. А. С. Попова» (СПбНТОРЭС). – Санкт-Петербург : СПбНТОРЭС, 2020. – № 75. – С. 155–156.