Abstract: Aiming to estimate for missing data of the electric energy metering and management system, the theory of optimum weighted combination which combined GM(1,1),SVM(Support Vector Machines) and ANN are introduced to shortterm load forecasting of the system, then the predicted value is get. The forecasting results based on the data from the distribution electricity center in Eastern Slovakia shows that the algorithm possesses evident effectiveness in the field of short-term load forecasting.

Key words: short-term load forecasting, GM(1, 1), SVM, ANN