Electricity demand forecasting of Rajshahi city in Bangladesh using fuzzy linear regression model

In this paper fuzzy linear regression method have been used to reduce the forecasting error. Traditional methods of forecasting (e.g. simple moving average, single exponential smoothing, regression analysis etc.) consider only past historical data. On the other hand this method takes into consideration the parameters impacting the demand along with past data. Two parameters-temperature and number of annual consumer has been considered in this paper. The method finds a minimum forecasting range (lower and upper bound) for each year based on past data and parameters. The parameters (temperature & number of annual consumer) has been forecasted using the autoregressive model. Mean absolute percentage error (MAPE) criteria has been used to measure the forecast accuracy.