AbstractPrevious research into computer-assisted methods of residential property valuation has concentrated upon statistical techniques i.e., multiple regression analysis.
Inquiries made of all leading academic and professional institutions in the United Kingdom, the United States of America, Australia and New Zealand involved in property valuation, indicate that this project is unique in current residential valuation research.
The. approach is based upon a study of the working practises of a professional mortgage valuer, his "expertise" and techniques utilised in the completion of mortgage valuations.A model of the valuation procedure has been developed and exposed to the critical evaluation of other valuers. This model has been implemented as a demonstration "expert system".
A critical evaluation of the suitability of the different software, knowledge elicitation and knowledge representation techniques for valuation work has been carried out and an assessment of the nature and use of uncertainty within the domain of mortgage valuation made.
The current methodology effectively demonstrates the knowledge-based concept of a separation of comparable property data, and the procedures used to manipulate that data, i.e. the valuer's use of comparable s. This enables the demonstration system to operate with few and "imperfect" comparables. Additionally the methodology is not time-related, the demonstration system selects comparable information from the same time period as the required valuation. These features are clearly an advance over the regression studies noted above which require complete data in large quantities over a restricted time period.
Currently the integration of knowledge-based and conventional data processing software is in its infancy and this is reflected in the limited nature of the demonstration system.
In conclusion the project has developed a wholly original approach to the problem of computer-assisted residential property valuation, contributing significantly to the available literature in the comparative method of valuation, computer-assisted valuation techniques, and the identification of uncertainty within a domain of expertise.
|Date of Award||Nov 1998|
|Supervisor||Stuart Gronow (Supervisor)|