MODELLING AND FORECASTING THE INTERNET AND MOBILE SUBSCRIBERS IN BANGLADESH

A TIME SERIES SEASONAL ARIMA APPROACH

Abstract

The internet and mobile devices are considered the backbone of today’s world. A country can only be called a digital country when it becomes an e-state and all of its activities are carried out over the internet. The main objective of this study is to build the best seasonal ARIMA model to forecast the internet and mobile users in Bangladesh. This study takes into account monthly data from the Bangladesh Telecommunication Regulatory Commission (BTRC) from January 2012 to December 2021. This study also used data from January 2022 to December 2022 to validate the model. The Box-Jenkins approach was applied to build the model as well as to forecast the internet and mobile subscribers. The findings show that SARIMA (1, 1, 1) (1, 0, 0)12 and SARIMA (1, 1, 0) (1, 0, 0)12 are the best models to forecast the internet and mobile subscribers, respectively, in terms of the lowest AIC, AIC_C, and BIC value. The value was forecasted with a 95% confidence interval using the best-fitted models for both internet and mobile subscribers. The original and forecasted series were compared, and the findings reveal that the selected models provide satisfactory results. The proposed models may help policymakers forecast accurately and know the future trends of internet and mobile users in Bangladesh.