Computing Uncertainty of Natural Neighbour Interpolation
conference contributionposted on 16.09.2019 by Thomas Etherington
Any type of content contributed to an academic conference, such as papers, presentations, lectures or proceedings.
Interpolation techniques provide a method to convert point sampling data into a continuous estimate of a field phenomenon and have become a fundamental geocomputational technique of spatial and geographical analysts. Natural neighbour interpolation is one method of interpolation that has several useful properties: it is an exact interpolator, it creates a smooth surface free of any discontinuities, it is a local method, and is spatially adaptive. However, as with any interpolation method, there will be uncertainty in how well the interpolated field values reflect actual phenomenon values. Using a method based on distance error rates calculated for data points via leave-one-out cross-validation, reasonable estimates of interpolation error can be made, at least within the convex hull of the data points. While this method does not replace the need for analysts to use sound judgement in their interpolations, it does provide a valuable tool to aid in assessing the uncertainty associated with those interpolations.