基于AEPSO改进支持向量机的堤基沉降预测

Application of an improved support vector machine based on AEPSO to forecast the settlement of embankment

在支持向量机的预测模型中,关键参数的选取是最重要的一步。在此基础上提出基于AEPSO改进的支持向量机预测模型。利用AEPSO的局部和全局搜索能力,提高支持向量机关键参数寻优的精度。文中详细介绍模型建立的过程,以莆田市木兰溪防洪工程为例,应用改进的模型进行堤基沉降预测,并与标准PSO支持向量机预测模型相对比。结果表明,基于AEPSO改进的支持向量机预测模型提高了预测的精度。

In support vector machine forecasting model, the selection of key parameters is the most important step. By using the AEPSO local and global search ability, the new support vector machine optimized is proposed. In order to verify the optimized effect, this paper introduces the process of establishing a model detailedly and takes an example of the Mulan Stream flood control project in Putian City. The results compared with PSO-SVM show the improved support vector machine forecasting model will improve the precision of prediction.