Broadband user demand forecasting in Indonesia based on Fourier analysis

Abstract

In this paper, a broadband user demand forecasting in Indonesia based on Fourier analysis is presented. In this approach, the historical data of broadband user demand is considered as time-series data. To obtain trend of the time-series data, polynomial expression is incorporated in the Fourier series representation instead of constant. The presented forecasting model is then confirmed its validity by using hypotethical daily price fluctuation data. It can be shown that the presented forecasting model gives good agreement between forecasted data and actual data. The same forecasting model then is applied to forecast broadband user demand in Indonesia. Several cases of fixed and mobile broadband are presented. It can be shown that the forecasting model based on Fourier analysis gives good result with the error of the estimated value is less than 10%.