论文部分内容阅读
为进一步提高灰色Verhulst模型的预测精度,将LS-SVM算法与灰色Verhulst模型相结合,对灰色Verhulst模型的参数估计方法和预测方法进行了改进。该方法采用LS-SVM算法,构造以背景值序列和原始序列为训练样本的LS-SVM,将Verhulst模型参数的估计问题转化为灰色LS-SVM的参数估计问题,依据LS-SVM算法求得灰色LS-SVM的参数,进而得到Verhulst模型的参数估计,方法上遵循了结构风险最小化原则,适合Verhulst小样本建模的特点。将改进的模型应用于软土地基建筑物的沉降预测,结果表明本文的方法是可行的且有效的,比传统方法预测精度高。
In order to further improve the prediction accuracy of gray Verhulst model, LS-SVM algorithm and gray Verhulst model are combined to improve the method and prediction method of gray Verhulst model. This method uses LS-SVM algorithm to construct LS-SVM with background value sequence and original sequence as training samples, and transforms the estimation problem of Verhulst model parameter into the parameter estimation problem of gray LS-SVM. According to the LS-SVM algorithm, LS-SVM parameters, and then get the Verhulst model parameter estimation, the method follows the principle of structural risk minimization, suitable for the characteristics of Verhulst small sample modeling. The improved model is applied to the settlement prediction of soft ground buildings. The results show that the proposed method is feasible and effective, and the prediction accuracy is higher than the traditional method.