中文摘要：

为解决大坝非线性变形数据中长期预报问题，提出一种新的模型即混沌-支持向量机模型：首先根据混沌相空间重构理论，对其进行相空间重构；然后根据统计学习理论，对融合外界影响因素的重构相空间里的变形数据，应用最小二乘支持向量机进行建模并做中长期预测。实际算例证实其应用效果较好。

英文摘要：

A new model （chaotic time series -least square support） is put forward for solving the problem of medium - long time prediction by use of the dam＇s nonlinear deformation data affected by the outer factors. Firstly, the nonlinear deformation data＇ s phase space is reconstructed based on the phase space reconstruction theory. Then LSSVM （Least Square Support Vector Machine） is used to model and make the medium-long time prediction from the data in the reconstruction phase space integrated with the outer factors based on the statistics-learning theory. An example testifies the model is effective.