论文部分内容阅读
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 time-series support vector machine model of tunnel surrounding rock deformation prediction is established. In order to solve the problem of parameter selection of support vector machine (SVM), the cultural fish swarm algorithm (CAAF) was used to search the related parameters of SVM, which avoided the blindness of artificial search parameters and improved the promotion performance of the model. Firstly, the sample data of tunnel surrounding rock deformation in the example are processed by time series to construct the learning and forecasting samples, and then the model is predicted by the support vector machine model of cultural fish swarm optimization. Compared with the BP neural network prediction results, this method has the advantages of fast calculation speed and high prediction accuracy, and has higher applicability to practical projects.