Evaluating Time Series Approaches for Cellular Network Traffic Forecasting

This paper examines the predictive accuracy of four distinct models-ARIMA, ARMA, AR, and Linear Regression-applied to network traffic forecasting. Through a comprehensive analysis, the ARIMA model was identified as the most effective for non-stationary data, demonstrating its robustness in capturing complex temporal patterns. The linear regression model, however, outperformed the others across the dataset, with the lowest Mean Squared Error (MSE) of 1.223 and Root Mean Squared Error (RMSE) of 1.106, indicating minimal prediction errors. The results underscore the importance of selecting appropriate models based on data characteristics and prediction requirements. Furthermore, the study highlights the potential of linear regression, a relatively simple method, to outperform more complex time series models under specific conditions.