||Bootstrapping. The neglected approach to uncertainty
||The history of confidence intervals is discussed together with the pedagogical predilection to a parametric view of the world. The weakness of these methods is that they are not related to any measurement of probability. The inappropriateness of such an outlook in the realm of real estate is illustrated with real world examples. A non-parametric vision is developed and the utility of the jackknife and bootstrap methods (Efron 1979) developed. Their similarity to Monte Carlo Simulation is demonstrated. The standard algorithm is derived and presented in both VBA code suitable for incorporation into Excel and in the C# language that is compatible with the .Net framework. An original recode that is an order of magnitude faster is also presented. Applications to confidence intervals for residential price indices and weighted multiple linear regression are demonstrated.
|Year of publication:
||bootstrap, monte carlo, uncertainty, confidence interval
Kershaw, Paul (2011).
Bootstrapping. The neglected approach to uncertainty. 18th Annual European Real Estate Society Conference in Eindhoven, the Netherlands,