机构地区: 北京工业大学建筑工程学院工程抗震与结构诊治北京市重点实验室
出 处: 《钢结构》 2007年第2期81-84,共4页
摘 要: 在介绍支持向量机的理论基础及其在时间序列预测中应用方法的基础上,建立一种基于支持向量机的结构变形预测方法。通过在某会展中心张弦桁架实际变形数据中的应用并与AR模型预测方法进行对比,验证了基于支持向量机的结构变形预测方法的准确性和优越性,该方法有较高的预测精度,并且在预测区间较长情况下同样有效,具有较大的研究意义并可以在实际预测中应用。 The theory basis of support vector machine (SVM) is introduced and SVM is applied into the research on time series forecast. A structural deformation forecast method based on SVM is established. This method is applied in the deformation forecast for the prestressed beam string structures of an international conference and exhibition center. The method is compared with autoregressive (AR) model, and the accuracy and advantage are verified. Structural deformation forecast based on SVM is effective for the case of forecast of long interval; it has positive research significance and could be applied in actual engineering.
关 键 词: 张弦桁架 支持向量机 变形预测 模型 时间序列预测
领 域: [建筑科学]