Demand Forecasting on Telephone Usage Based Kalman Filtering

This paper presents a method of demand forecasting on telephone usage which is based on Kalman filtering. The Kalman filter is a method that provides an efficient computational solution in least square sense. It is thus received much attention for estimation purpose. In order to visualize the efficiency of the proposed technique, it is used to predict the telephone usage demand of Thailand nationwide and only in Bangkok. The data employed in the forecasting process are the former forecasted demand and the number of population during 1954-2003 (for Bangkok area) and 1991-2010 (for nationwide). The result of this method is compared with that of statistical analysis method. It is found that the proposed technique is superior to statistical methods. Furthermore, it employs a simple mathematical process with less computational consumption