Conformal Based Uncertainty Bands for Predictions in Functional Ordinary Kriging
摘要
Functional Ordinary Kriging is the most widley used method to predict a curve at a given spatial point by computing a weighted average of observed functional values in space. Determining the uncertainty associated with predicted curves remains a challenge. To address this issue we introduce a distribution free prediction method for the proposed model and illustrate its performances using the well known data set of Canadian Temperature [12].