Short - term electric load forecasting using SVR implementing LibSVM package and Python code

Electrical load forecasting is an important topic within the electrical market which has been done by a machine learning methodology: Support Vector Machines (SVM). Load forecasting with SVM will form the non-linear relations with the parameters that have an effect on the load; additionally to the correct modeling of the load curve on weekends and holidays. The past information is used as a sample for the applying and therefore holidays associated demand as an important factor inprediction. The LibSVM package and Python codeis used for modeling the SVM. Resultsare obtainedand comparison is made for the two methods.