Short-term load forecasting using least squares support vector machine / Abdul Zamer Afiq Abd Razak

In the electrical industry, accurate forecast of electricity load has been highlighted as most of the important issues. This paper proposes a model for short-term load forecasting using least-square support vector machines. The collected data are from Dayton, Ohio, United State. This collected data are analyzed and suitable features are selected for the model. Last 24 hour load demands are used to the features of load forecasting in combination with days of the week and hours of the day. The suitable data set is used for the model training, and then forecasting of day ahead hourly load demands is performed. The experimental results, obtained from a real-life benchmarks, showing that the proposed model is effective and accurate.