This research, using valid sales of single-family homes in Norfolk, Virginia, from 2010 to 2012, evaluated the varying predictability power that different kernel and bandwidth specifications in geographically weighted regression models lend to mass appraisal of real estate, and the potential improvement each lends to taxing entities in attaining equity, uniformity, and ultimately defensibility in their property assessments. Specifically the weighting specifications that were studied were the Gaussian kernel with fixed bandwidth, the Gaussian kernel with adaptive bandwidth, the bi-square kernel with fixed bandwidth, and the bisquare kernel with adaptive bandwidth. The model applying a Gaussian kernel and adaptive bandwidth produced results that were most uniform by IAAO standards.
Multiple regression analysis
Bidanset, P. E., & Lombard, J. R. (2014). The effect of kernel and bandwidth specification in geographically weighted regression models on the accuracy and uniformity of mass real estate appraisal. Journal of Property Tax Assessment & Administration, 11(3), 5-14. Retrieved from https://researchexchange.iaao.org/jptaa/vol11/iss3/1