Predicting electricity consumption: A comparative analysis of the accuracy of various computational techniques

This research explores the dynamic relation between price, temperature and humidity; and its effect on electricity consumption of electric appliances. It develops prediction models for electricity consumption based on these variables. It is important that reliable methods are employed in modelling and prediction of energy needs otherwise inappropriate models and poor forecasts may occur. In this research, prediction estimates for the daily electricity consumption for a local university in Malaysia was computed using regression model, artificial neural network (ANN) and the kalman filter adaptation algorithm. The estimates of the methods were compared using performance measures based on statistical parameters obtained from identifying the difference between actual and predicted values. This research identified the kalman filter adaptation algorithm as the bests performing method in making predictions for future electricity consumption.