摘要
在海量数据中,建立准确的电价预测模型,对于企业和电力用户制定合理的决策具有重要意义.针对影响电价预测模型的数据量较大的问题,采用主成分分析(PCA)方法提取主要特征,降低数据维度.为了提高电价预测的准确度,考虑到传统狼群算法随机初始化和线性收敛因子影响收敛速度和收敛精度的问题,提出佳点集初始化种群和双曲收敛因子方法,并采用改进的狼群算法优化核极限学习机的正则化系数C和核参数g,以提升核极限学习机的稳定性和泛化能力.仿真结果表明,改进的灰狼算法具有更优的收敛速度和收敛精度,而且改进的灰狼优化的核极限学习机相比于传统的算法更适用于电价预测.
Abstract
Establishing an accurate electricity price prediction model in the mass data is of great significance for enterprises and power users to make reasonable decisions.In view of the large amount of data affecting the electricity price prediction model,principal component analysis(PCA)method is adopted to extract the main features and reduce the data dimension.In order to improve the accuracy of electricity price prediction,the good point set initialization population and the hyperbolic convergence factor method are proposed by considering that the random initialization and linear convergence factor of traditional wolf pack algorithm affect the convergence speed and accuracy.In order to improve the stability and generalization ability of the kernel ex-treme learning machine,the regularization coefficient C and kernel parameter g of the kernel extreme learning machine are opti-mized by the improved wolf pack algorithm.Simulation results show that the improved gray wolf algorithm has better conver-gence speed and accuracy.And the improved gray wolf optimized kernel extreme learning machine is more suitable for electrici-ty price prediction than the traditional algorithm.
维普
万方数据