Recursive Gaussian process: On-line regression and learning
Recursive Gaussian process: On-line regression and learning
复制标题
DOI：
10.1016/j.patrec.2014.03.004
复制
发表时间：
2014-08-01
影响因子：
5.1
通讯作者：
Huber, Marco F.
中科院分区：
文献类型：
--
作者：
Huber, Marco F.
提出了两种具有较低计算和存储需求的在线高斯过程回归方法。第一种方法假定已知的超参数，并在存储潜在函数的均值和协方差估计的一组基础向量上执行回归。第二种方法是在线学习超参数。为此，利用了非线性高斯状态估计技术。将所提出的方法与最新的稀疏高斯过程算法进行了比较。(C)2014爱思唯尔B.V.保留所有权利。
Two approaches for on-line Gaussian process regression with low computational and memory demands are proposed. The first approach assumes known hyperparameters and performs regression on a set of basis vectors that stores mean and covariance estimates of the latent function. The second approach additionally learns the hyperparameters on-line. For this purpose, techniques from nonlinear Gaussian state estimation are exploited. The proposed approaches are compared to state-of-the-art sparse Gaussian process algorithms. (C) 2014 Elsevier B.V. All rights reserved.