TY - GEN
T1 - Adding Value to Desktop Valuations: The Use of Artificial Intelligence Techniques to Assist in Residential Property Price Forecasting
AU - Wilson, Ian
AU - Ware, Andrew
AU - Jenkins, David
AU - Paris, Stuart
PY - 2001
Y1 - 2001
N2 - Property values are currently assessed by professional valuers, who estimate values based on current bid prices (open market values). This approach has failed to predict the periodic market crises or to produce estimates of long-term sustainable value. This research examines the use of artificial intelligence techniques, trained using national economic, social and residential property transaction time-series data, to forecast trends within the housing market. Work was undertaken to identify a measure of house prices that showed no overall trend. Such a representation has two major advantages. First is that the measure represents a time independent indicator of the state of the housing market in relation to its underlying average. Second is that it is easier to isolate those factors causing the measure to vary over time, and it allows the removal of the date as an explicit variable in models. Neural networks were trained using as inputs economic data and the previous quarter’s measure of house prices. The single output was the current quarter’s measure of house prices. The trained networks were used to make successive one period (quarter) ahead forecasts. The results show that artificial neural networks, trained using national economic, social and residential property transaction time-series data, can be used to forecast trends within the housing market under various market conditions.
AB - Property values are currently assessed by professional valuers, who estimate values based on current bid prices (open market values). This approach has failed to predict the periodic market crises or to produce estimates of long-term sustainable value. This research examines the use of artificial intelligence techniques, trained using national economic, social and residential property transaction time-series data, to forecast trends within the housing market. Work was undertaken to identify a measure of house prices that showed no overall trend. Such a representation has two major advantages. First is that the measure represents a time independent indicator of the state of the housing market in relation to its underlying average. Second is that it is easier to isolate those factors causing the measure to vary over time, and it allows the removal of the date as an explicit variable in models. Neural networks were trained using as inputs economic data and the previous quarter’s measure of house prices. The single output was the current quarter’s measure of house prices. The trained networks were used to make successive one period (quarter) ahead forecasts. The results show that artificial neural networks, trained using national economic, social and residential property transaction time-series data, can be used to forecast trends within the housing market under various market conditions.
M3 - Conference contribution
BT - Proceedings of: The Cutting Edge 2001, The Real Estate Research Conference of the RICS Research Foundation
ER -