A class of multi-resolution approximations for large spatial datasets
A class of multi-resolution approximations for large spatial datasets
复制标题
DOI：
10.5705/ss.202018.0285
复制
发表时间：
2017-10
影响因子：
1.4
通讯作者：
M. Katzfuss;Wenlong Gong
中科院分区：
文献类型：
--
作者：
M. Katzfuss;Wenlong Gong
高斯过程是用于空间、时间和函数数据的流行且灵活的模型，但它们在计算上不适用于大型数据集。我们讨论高斯过程近似，它使用多分辨率的基函数来实现快速推理，并且可以（近似）表示任何协方差结构。我们认为这种多分辨率近似框架，锥形版本和域分区（块）版本的两种特殊情况。我们描述的理论性质和推理过程，并研究计算复杂性的方法。还提供了数值比较和卫星数据的应用。
Gaussian processes are popular and flexible models for spatial, temporal, and functional data, but they are computationally infeasible for large datasets. We discuss Gaussian-process approximations that use basis functions at multiple resolutions to achieve fast inference and that can (approximately) represent any covariance structure. We consider two special cases of this multi-resolution-approximation framework, a taper version and a domain-partitioning (block) version. We describe theoretical properties and inference procedures, and study the computational complexity of the methods. Numerical comparisons and an application to satellite data are also provided.