摘要
电网的安全经济可靠运行需要电力负荷预测具有较高的精度。尽管支持向量机理论解决电力负荷预测数据小样本,非线性,局部极小点等问题有很大的优势,但支持向量机的参数(c,σ)难以确定最优值。采用带惯性权重的粒子群优化算法(PSO)对支持向量机参数寻优并进行电网短期负荷预测。将预测结果同普通支持向量机和RBF神经网络预测结果对比,结果证明这种方法减少了预测耗时,提高了预测的稳定性和精度。
The efficient operation of the grid needs power load forecasting has higher precision. Support Vector Machine have a great advantage in solving the problem of small data sample, non-linear, local minimum points, so we can use the theory of support vector regression to predict the load. We use the particle swarm optimization （PSO） to search for the optimal parameters of support vector machine because it is difficult to determine them. The predicted results were compared with ordinary Support Vector Machine model and RBF neural network model. The comparison shows that this method can improve the prediction accuracy and stability and reduce the running time.
关键词
支持向量回归
粒子群优化
电力负荷预测
Support Vector Regression
Particle Swarm Optimization
Power Load Forecasting
