摘要
为了降低人为设定参数值对支持向量机运行结果准确度的影响,采用智能算法中的人工鱼群算法,搜寻支持向量机相应参数的最优解。由于人工鱼群算法运算在寻优精度和效率方面均有提升空间,故将混沌机制引入人工鱼参数初始化,通过改进固定参数和行为算子得到支持向量机预测模型。使用该模型进行中长期电力需求预测研究,并与其它参数优化算法产生的模型进行均方误差对比。研究结果表明:改进后的模型在拟合均方误差和预测均方误差上都优于未优化的模型,支持向量机在预测精度方面有一定程度的提升。
In order to weaken the effect that artificial parameter values influence on accuracy of result of support vector machine （SVM）, this paper uses the artificial fish swarm algorithm（AFSA）, one of the intelligence algorithms, to search the optimum solution of corresponding parameters of SVM. Owing to the room for improvement in accuracy and efficiency of AFSA, the cha- os mechanism is introduced into the parameter initialization of artificial fishes with the improvement of fixed parameters and op- erator, which reaches the prediction model of SVM based on parameter optimized by improved AFSA, and helps the prediction study of mid-and-long term power demand, which compares with models optimized by other algorithm. The results show that the improved model is more excellent than the model without optimization in mean square error （MSE） of fitting and forecas- ting. Besides that, the prediction accuracy of SVM has a certain degree of improvement.
出处
《软件导刊》
2018年第3期183-186,共4页
Software Guide
关键词
中长期负荷预测
支持向量机
人工鱼群算法
混沌机制
参数优化
均方误差
WEN Jue （School of Optoelectronic Information and Computer Engineering, University of Shanghai for Science and Technology, Shanghai 200093, China）
作者简介
温珏（1993-），女，上海理工大学光电信息与计算机工程学院硕士研究生，研究方向为图像处理、模式识别。
