论文部分内容阅读
讨论了嵌入维数d和时间延迟?作为空间重构参数对LS-SVM预测模型精度的影响,提出了基于PSO参数优化的LS-SVM预测方法。将d、?以及模型参数(正则化参数?、核函数宽度?)作为优化对象,利用PSO方法对4个参数共同优化选取,建立LS-SVM风速预测模型。对2组风速数据进行了实验研究,结果显示该方法预测误差约为5.79%和7.33%。而对比方法 (单纯优化?、?)的误差为8.22%和11.10%。这一结果表明,同时对d、?、?、?进行优化选取是有必要的,相对于单纯优化?、?的模型,该方法可以大大提高预测模型精度。
The influence of spatial reconstruction parameters on the accuracy of LS-SVM prediction model is discussed, and LS-SVM prediction method based on PSO parameter optimization is proposed. Taking d, and model parameters (regularization parameter and kernel function?) As optimization objects, four parameters are jointly and optimally selected by using PSO method to establish LS-SVM wind speed prediction model. Two groups of wind speed data were studied experimentally, the results show that the prediction error of this method is about 5.79% and 7.33%. The error of the comparison method (pure optimization?,?) Is 8.22% and 11.10% respectively. This result shows that it is necessary to optimize d,?,?,? At the same time. Compared with the simple optimization model, this method can greatly improve the accuracy of the prediction model.