'BELIEF' in the past: Dempster-Shafer theory, GIS and archaeological predictive modelling

This paper introduces the use of a new technique in archaeological predictive modelling which has particularly wide ranging appeal and application in the cultural heritage management sphere. Most predicitve modelling programs are restricted to the use of certain types of data with a range of untested assumptions or caveats. Traditional probability based models are difficult to construct, and are often limited by the use of data sets which render the results statistically suspect. The method introduced here makes use of non-parametric statistical methods which are not hampered by imperfect raw data. In particular, this paper introduces the use of Dempster-Shafer theory in archaeological predictive modelling.