摘要
建立了隧道围岩变形预测的基于时间序列的支持向量机模型。针对支持向量机(SVM)的参数选择问题,运用文化鱼群算法(CAAF)来搜索支持向量机的相关参数,避免了人工搜索参数的盲目性,提高了模型的推广性能。该模型首先将实例中隧道围岩变形的样本数据进行时间序列处理,构建学习和预测样本,再利用文化鱼群优化的支持向量机模型进行预测。与BP神经网络预测结果相比,该方法运算速度快、预测精度高,对实际工程具有更高的适用性。
A tunnel surrounding rock deformation prediction based on support vector machine(SVM)model of time series is established.Because of the support vector machine(SVM)parameters issue,fish culture algorithm(CAAF)is used to search for SVM parameters to avoid the blindness of manual searching and to improve the promotion of the model.At first the tunnel surrounding rock deformation data in the instance is processed by time series,and build learning and prediction data.Then the support vector machine(SVM)model optimized by culture fish algorithm(CAAF)is applied to forecast.Compared with the BP neural network prediction results,the method has high computing speed,high precision,and has higher applicability to the practical engineering.
出处
《公路》
北大核心
2016年第3期221-225,共5页
Highway
基金
国家自然科学基金项目
项目编号51274053
关键词
隧道围岩变形
时间序列处理
文化鱼群算法
支持向量机
参数选择
tunnel surrounding rock deformation
time series
culture and fish algorithm
support vector machine(SVM)
parameter selection