Artificial Intelligence (AI) and machine learning (ML) algorithms have proved to be game-changers in the field of traffic forecasting for fifth-generation (5G) and sixth-generation (6G) networks. Researchers at RUDN University have recently conducted a study to explore the effectiveness of AI and ML in predicting mobile network profiles. By leveraging these advanced technologies, network providers can better plan and manage network traffic, leading to improved user satisfaction and network efficiency.
In their study, the researchers focused on two popular time-series analysis models: the Holt-Winter model and the Seasonal Integrated Autoregressive Moving Average (SARIMA). Using a dataset from a Portuguese mobile operator, the researchers aggregated hourly traffic statistics to train and test the models. They found that both models performed exceptionally well in forecasting traffic within the next hour.
The SARIMA model demonstrated its strength in predicting user-to-base station traffic, achieving an average error rate of just 11.2%. This model excels at monitoring transient patterns in mobile network traffic due to its capacity to record temporal patterns. On the other hand, the Holt-Winter model showed better performance in estimating base station-to-user traffic, with an error rate of up to 4%. This model’s ability to handle intricate seasonality and trend components contributed to its accuracy.
To measure the models’ performance, the researchers used various criteria such as Mean Square Error (MSE), Root Mean Square Error (RMSE), Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), and Mean Scaled Logarithmic Error (MSLE). While the models achieved impressive results, the researchers highlighted the potential for further improvement through fine-tuning specific hyperparameters.
The researchers emphasized the need for combining statistical models with AI and ML techniques to enhance the accuracy of traffic predictions and promptly detect anomalies. As the researchers continue to explore methods to optimize performance and enhance user satisfaction, this study holds significant implications for the efficiency of 5G and 6G networks.
With the introduction of cutting-edge technology and a relentless pursuit of accuracy in forecasting network traffic, AI and ML algorithms bring new possibilities to the world of telecommunications. As providers strive to maximize network efficiency, the invaluable insights gained from this research will drive advancements in this rapidly evolving field.
The source of the article is from the blog reporterosdelsur.com.mx