Buying a house is the biggest financial gamble most people will ever make.
A team of academics hopes to take some of the guesswork out of the process by making an advanced house price forecast service freely available on the internet.
The system, devised at the University of Glamorgan, analyses government statistics to make its predictions of how prices might change over the next three years.
This will give potential customers assistance when it comes to deciding whether a particular property is affordable.
David Jenkins, reader in land economy, said that many valuers used instinct rather than analysis of trends to work out a market price. Institutions that have devised forecasting methods tend to simply extrapolate on trends in open-market values.
The Glamorgan system, which will shortly be made available online, is based on the concept of a property's sustainable value, which Mr Jenkins hopes will be more useful for the purposes of predicting the future, especially given the distorting effects of the recent boom.
It uses government forecasts of eight different trends that affect property prices. These include average earnings, bank rates and spending figures.
Ian Wilson, an operations research fellow, said a neural net was fed a large amount of data about previous values of these indicators and the prices that resulted.
This created a model that could then be fed the latest government predictions to come up with forecasted prices for the next three years.
The system operates at a regional level and does not yet work at a city or neighbourhood level of resolution.
The team believes it will still prove useful, especially given that most problems with mortgages emerge in the first three years.
"This will give the borrower a feel for what they are letting themselves in for based on what the official figures suggest is going to happen in the housing market," Mr Jenkins said.
It will also incorporate "what if" scenarios that deal with changes in interest rates.
A report on the research is under review at the Engineering and Physical Sciences Research Council.