摘要
针对变形监测数据的随机性和非平稳性,以及单一预测模型的不足,该文提出了基于小波去噪的灰色最小二乘支持向量机变形预测模型。采用小波去噪对原始数据进行降噪处理,减弱数据随机扰动的影响,建立灰色最小二乘支持向量机模型,并采用粒子群算法寻找最优参数。通过大坝位移监测数据实例对模型进行验证,并与灰色模型、最小二乘支持向量机以及灰色最小二乘支持向量机进行对比分析。实验结果证明,该模型预测精度更高、稳定性更强。
Aiming at the randomness, non-stationary of deformation monitoring data and deficiency of single forecasting model, the model of a grey least square support vector machine based on wavelet denoising was presented in this paper. The wavelet denoising method was adapted to deal with the original data of deformation in order to weaken the effect of random disturbance factors. Then, the grey least square sup- port vector machine was established and the particle swarm optimization algorithm was used to find the op timal parameters. An example based on the measured data of a dam deformation was validated with this model. Grey model, least square support vector machine and grey least square support vector machine model were used to compare with the proposed model. The results proved that the presented model was more stable and the forecasting precision of the presented model was higher.
出处
《测绘科学》
CSCD
北大核心
2017年第10期134-137,160,共5页
Science of Surveying and Mapping
基金
国家自然科学基金项目(50604009)
辽宁省"百千万人才工程"人选资助项目(2010921099)
关键词
变形预测
小波去噪
灰色模型
最小二乘支持向量机
粒子群算法
deformation forecast
wavelet denoising
grey model
least square support vector ma-chine
particle swarm optimization
