基于人工免疫加权支持向量机的电力负荷预测

提出了一种人工免疫加权支持向量机负荷预测模型,针对各训练样本重要性的差异,提出了给各个样本的参数赋予不同权重的加权支持向量机方法,并用人工免疫算法对支持向量机的核函数和参数进行寻优,从而很好的解决支持向量机应用中核函数和参数选择这一公认的难题,减少了人工凭经验选择的盲目性。经过仿真,证明了其在短期负荷预测中的有效性。

Power load forecasting based on artificial immune algorithm-weighted-SVM model

A new artificial immune algorithm weighted support vector machine forecasting model is presented in this paper, Different training example has different importance, so the author gives different weight to the parameter of different training example. And artificial immune algorithm to select the most suitable kernel function and parameter are also employed. It is well known that there are few theories about how to select the kernel function and parameter, the problem can be solved by this way rather than by experience. Simulation results have proved its validity and effectiveness.