Stochastic Gradient Algorithms for AR Models with Missing Data
Stochastic Gradient Algorithms for AR Models with Missing Data
复制标题
DOI：
--
复制
发表时间：
2008
期刊：
影响因子：
--
通讯作者：
Ding Feng
中科院分区：
文献类型：
--
作者：
Ding Feng
利用多项式变换技术，将含有缺失观测数据的自回归模型转化为可从稀缺观测数据中识别的特殊模型，并利用模型等价原理和基于残差的随机梯度算法估计缺失数据。最后给出了一个仿真算例。
By using the polynomial transform technique, the auto-regression model with missing observation data is transformed into a special model which can be identified from scarece observation data, and the missing data are estimated by the model equivalence principle and a residual based stochastic gradient algorithm. A simulation example is included.