Abstract

Traditionally, archaeological predictive models have used all sites in a region and all time periods to make predictions about where archaeological sites might be located. This approach has come under heavy criticism, given the assumption that sites would have been located according to different criteria on both functional and temporal lines. While it is sometimes difficult to ascertain site function, site temporality is much less problematic.This paper examines the initial results of a project to do archaeological predictive models based on time slices, in order to improve predictive abilities of models. It will examine the difficulties, both methodological and practical of taking this type of approach, and offer tentative conclusions about the benefits of including a temporal dimension in the modeling process.