作者: W.J. McCluskey , M. McCord , P.T. Davis , M. Haran , D. McIlhatton
DOI: 10.1080/09599916.2013.781204
关键词:
摘要: The advancement of computational software within the last decade has facilitated enhanced uptake mass appraisal methodologies by valuation and prediction accuracy in computer-assisted community for price modelling, estimation tribunal defence. Applying a sample 2694 residential properties, this paper assesses analyses number geostatistical approaches relative to an artificial neural network (ANN) model traditional linear hedonic pricing purposes. findings demonstrate that localised regression approach is superior terms explanation, reliability accuracy. ANNs can be shown perform very well predictive power, therefore accuracy, outperforming multiple analysis (MRA) approaching performance spatially weighted approaches. However, retain ‘black box’ architecture limits their...