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 general, stepping-regression techniques. The problem of colinearity between the regressors is investigated by using standard statistical procedures. Model adequacy is determined from a residual analysis technique. The final model involves the parameters air temperature, relative humidity, solar radiation, and population. However, the model coefficients are functions only of solar radiation and relative humidity or constant.