A REGRESSION-MODEL FOR ELECTRIC-ENERGY-CONSUMPTION FORECASTING IN EASTERN SAUDI-ARABIA

Electrical energy consumption in Eastern Saudi Arabia is modeled as a function of weather data, global solar radiation and population. Five years of data have been used to develop the energy-consumption model. Variable selection in the regression model is carried out by using gen eral, stepping-regression techniques. The problem of colinearity betwe en the regressors is investigated by using standard statistical proced ures. Model adequacy is determined from a residual analysis technique. The final model involves the parameters air temperature, relative hum idity, solar radiation, and population. However, the model coefficient s are functions only of solar radiation and relative humidity or const ant.