Measuring the VC-dimension of a learning machine
Neural Computation
Neural Networks: A Comprehensive Foundation
Neural Networks: A Comprehensive Foundation
Comparison among five evolutionary-based optimization algorithms
Advanced Engineering Informatics
An improved GA and a novel PSO-GA-based hybrid algorithm
Information Processing Letters
Hi-index | 0.00 |
Accurate short-term load forecasting is important for performing many power utility functions, including generator unit commitment, hydro-thermal coordination and so on. Power load forecasting is complex to conduct due to its nonlinearity of influenced factors. According to the chaotic and non-linear characters analyze of power load data and the theory of phase-space reconstruction, the model of support vector machines based on Lyapunov exponents was established. The time series matrix was established, and then Lyapunov exponents were computed to determine time delay and embedding dimension. A new incorporated intelligence algorithm is proposed and used to determine free parameters of support vector machines in order to improve the accuracy of the forecasting. Subsequently, power load data of Inner Mongolia Autonomous Region are employed to verify the new model. The empirical results reveal that the proposed model outperforms the SVM model. The results show that the presented method is feasible and effective.