The Uncertain Canonical Process Regression Models

摘要：

In this paper, we investigate the regression models with error terms that are the members of an uncertain canonical process. This new model may be regarded an extension to the Bayesian regression models with a Gaussian process as variance-covariance prior, named as Gaussian process regression. The variancecovariance matrix is no longer an identity matrix (with positive constant multiplier), but it is intrinsic to the uncertain canonical process, which results in a weighted repression model.

关键词： Uncertain measure Gaussian processes uncertain canonical process intrinsic autocovariance matrix weighted regression model

作者: Renkuan Guo Wei Dai Danni Guo Yanhong Cui Tim Dunne

作者单位: University of Cape Town, Private Bag, Rondebosch 7701, Cape Town, South Africa Central University of Finance and Economics, Beijing, China South African National Biodiversity Institute, Kirstenbosch, Cape Town. South Africa

会议类型: 国际会议

会议名称: The Second International Conference on Uncertainty Theory(ICUT)(第二届不确定理论国际会议)

会议地点: 拉萨

会议语种:英文

页码: 36-44

在线出版日期: 2011-08-06（万方平台首次上网日期，不代表论文的发表时间）