Output details

Revisiting Kappa to account for change in the accuracy assessment of land-use change models

In recent years there has been considerable academic debate over the validity of the most commonly used metric for quantifying spatial simulation goodness-of-fit: the Kappa statistic. Instead of presenting further arguments to the discussion, this paper presents a solution: a new metric called Kappa Simulation that incorporates the attractive qualities lauded by the Kappa proponents and avoids the problems raised by Kappa opponents. This paper influenced the course of the academic debate, and recent studies are using the metric as part of validation and calibration routines (e.g. Koch et al. 2012, Wang & Marceau 2013, Soares-Filho et al. 2013).