Prediction performance of a hedonic pricing model for housing

Recent studies point to the superiority of the prediction ability of semiparametric methods over conventional, parametric hedonic price functions for housing. This study evaluates the predictive performance of three parametric forms-linear, quadratic, and cubic-and uses an index of location to incor...

Full description

Saved in:
Bibliographic Details
Published in:The Appraisal journal 2005-03, Vol.73 (2), p.198
Main Authors: Laurice, Jennifer, Bhattacharya, Radha
Format: Article
Language:English
Subjects:
Appraisals
Automation
Censuses
Dependent variables
Dwellings
Forecasts and trends
Housing
Management
Mathematical models
Methods
Mortgage rates
Neighborhoods
Prices
Pricing
Statistics
Studies
Variables
Online Access:Get full text
Tags: Add Tag
No Tags, Be the first to tag this record!
Description
Summary:Recent studies point to the superiority of the prediction ability of semiparametric methods over conventional, parametric hedonic price functions for housing. This study evaluates the predictive performance of three parametric forms-linear, quadratic, and cubic-and uses an index of location to incorporate neighborhood effects on sale prices. For the cubic model, the percentage mean absolute prediction errors range from 11.43% to 14.89% in the holdout sample. The model presented performs reasonably well and is computationally simpler than semiparametric methods, reaffirming that an augmented parametric model is an inexpensive tool that can serve as a valuable complement to the appraisal process. [PUBLICATION ABSTRACT]
ISSN:0003-7087