对振动测试数据评价参数的数学反演使用时间序列的线性和非线性模型进行了建模，计算结果表明非线性的ＳＥＴＡＲ模型描述测试数据所反映的物理过程比线性ＡＲＭＡＸ模型更趋于合理、全面和客观。非线性ＳＥＴＡＲ模型能较深刻描述被测体系的振动机理和物理行为，为下一步振动数据评判奠定了基础 Mathematical inversion of vibration test data evaluation parameters was modeled using linear and nonlinear models of time series. The calculation results show that the nonlinear SETAR model describes the physical process reflected by the test data is more reasonable and comprehensive than the linear ARMAX model. And objective. The nonlinear SETAR model can describe the vibration mechanism and physical behavior of the measured system more profoundly, laying the foundation for the next vibration data evaluation.