基于灰色关联和相关向量机的隧道地表沉降预测

发布时间：2018-05-03 15:42

本文选题：灰色关联 + 相关向量机　；参考：《大地测量与地球动力学》2017年10期

【摘要】：基于灰色关联算法确定与地表沉降有直接重要关联的主要影响因子,构建高斯核函数和多项式核函数的加权核函数,利用遗传算法优化模型参数,建立相关向量机地表沉降预测模型。实验结果表明,灰色关联算法能定量地反映系统影响因子与地表沉降变化的关联程度,有效处理不是完全明确的灰色系统信息;加权核函数的合理组合可较好地通过低维空间线性不可分映射变换到高维特征空间线性可分;遗传算法具有计算过程简单和自适应迭代寻优特点;相关向量机模型可极大地减少核函数的计算量,计算过程和结果均具有概率解释。该模型预测结果的多项精度指标值均优于BP神经网络和GR-SVM方法。
[Abstract]:Based on the grey correlation algorithm to determine the main influencing factors directly related to the surface subsidence, the weighted kernel functions of Gao Si kernel function and polynomial kernel function are constructed, and the parameters of the model are optimized by genetic algorithm. The prediction model of surface subsidence based on correlation vector machine is established. The experimental results show that the grey correlation algorithm can quantitatively reflect the degree of correlation between the influence factors of the system and the change of surface subsidence, and it is not completely clear about the grey system information. The reasonable combination of weighted kernel function can be transformed into high dimensional feature space by linear inseparability mapping in low dimensional space, and genetic algorithm has the characteristics of simple calculation and adaptive iterative optimization. The computation of kernel function can be greatly reduced by the model of correlation vector machine, and the calculation process and result have probability explanation. The prediction results of this model are superior to BP neural network and GR-SVM method.
【作者单位】：东华理工大学测绘工程学院;武汉大学测绘学院;
【基金】：国家自然科学基金(41374007) 江西省自然科学基金(20132BAB201049,20161BAB203102) 江西省教育厅科学技术研究项目(GJJ150555)~~
【分类号】：U456.31

【相似文献】