Journal of Real Estate Research · 1993 · 43 citations · 4 references
BusinessDemographyReal EstateHousingReal Estate EconomicsReal Estate ValuationHousing CharacteristicsHousing EconomicsProperty ValuationLot Size AnalysisLot Size PatternsMachine Learning In ValuationMultiple RegressionNeural Network ModelingNeural Network PredictionProperty Characteristics AnalysisProperty Value Appreciation
Empirical studies using multiple regression find the value of a residential property declines with its age. Because these results confirm the fact of physical deterioration of a house over time, little attention is paid to the statistical technique's inherent shortcomings. Accordingly, this paper uses a neural network, which is able to overcome multiple regression's methodological problems, to re-examine the effect of age on a house's value. We find that a negative relationship of value to age holds only for the first sixteen to twenty years of the life of a house. Then, not only does the decline in value stop, but a house actually starts to experience appreciation related, in part, to its lot size.
4
Multilayer feedforward networks are universal approximators
Kurt Hornik, Maxwell B. Stinchcombe, Halbert White · Neural Networks · 1989 · 21.2K citations