Constrained Kalman Filtering: Additional Results
Constrained Kalman Filtering: Additional Results
复制标题
DOI：
10.1111/j.1751-5823.2010.00098.x
复制
发表时间：
2010-08
影响因子：
2
通讯作者：
A. Pizzinga
中科院分区：
文献类型：
--
作者：
A. Pizzinga
本文研究基于一般状态向量线性约束的线性状态空间建模。讨论集中在四个主题：约束卡尔曼滤波与递归限制最小二乘估计；条件期望框架下约束卡尔曼滤波的新证明简化状态空间模型下的线性约束以及线性约束下的状态向量预测。所提出的技术在两个实际问题中得到了说明。第一个问题与动态因素模型下的投资分析有关，而第二个问题则是关于在GDP基准估计中进行约束预测。
This paper deals with linear state space modelling subject to general linear constraints on the state vector. The discussion concentrates on four topics: the constrained Kalman filtering versus the recursive restricted least squares estimator; a new proof of the constrained Kalman filtering under a conditional expectation framework; linear constraints under a reduced state space modelling; and state vector prediction under linear constraints. The techniques proposed are illustrated in two real problems. The first problem is related to investment analysis under a dynamic factor model, whereas the second is about making constrained predictions within a GDP benchmarking estimation.