Predictive Site Location
Predictive Site Location Modeling · Also known as: predictive modeling, maxent modeling
Predictive site location modeling uses machine learning algorithms (particularly maximum entropy models) to predict the probability of archaeological site occurrence across a landscape based on environmental and spatial variables. Developed for ecology but adapted for archaeology, predictive modeling identifies areas with high archaeological potential, guiding survey strategies and resource management.
Read the full method
Sign in with a free account to read this section.
Method map
The neighbourhood of related methods — select a node to explore.
When to use it
Apply predictive modeling to large regions when planning archaeological survey or assessing development impact. Particularly valuable for scoping survey in areas with incomplete site inventory. Works best when known sites are numerous (50+) and representative of all site types.
Strengths & limitations
- Objective, quantitative prediction of archaeological potential
- Guides survey efficiently to high-probability areas
- Identifies unexpected site-environment relationships
- Scalable to large regions
- Model accuracy depends on completeness of known site database—biased site samples produce biased predictions
- Environmental variables alone cannot explain all site location decisions—cultural factors unrepresented
- Over-reliance on models can bias survey toward predicted areas, missing sites in unpredicted locations
- Model results depend on variable selection and parameterization choices
Frequently asked
What is maximum entropy modeling and why is it useful for archaeology?
Maximum entropy finds the probability distribution most consistent with known site locations and environmental variables, without making assumptions about distribution shape. It works well with presence-only data (known sites but not systematic absence data), which is typical in archaeology.
How complete must the archaeological site inventory be for valid modeling?
The more complete, the better. If only 10% of sites are known, model predictions reflect only those discovered sites, biasing results toward easily found sites. Models trained on 30-50% of sites can yield useful results; models on smaller inventories are less reliable.
Can predictive models account for cultural preferences not captured by environmental variables?
Not directly. Models based only on environmental variables miss cultural factors. Combining environmental modeling with ethnographic or historical evidence about settlement preferences strengthens predictions.
Sources
- Phillips, S. J., Anderson, R. P., & Schapire, R. E. (2006). Maximum entropy modeling of species geographic distributions. Ecological Modelling, 190(3-4), 231-259. DOI: 10.1016/j.ecolmodel.2005.03.026 ↗
- Verhagen, P., & Whitley, T. W. (2012). Predictive modelling for archaeological resource management. Journal of Archaeological Science, 39(5), 1066-1077. link ↗
How to cite this page
ScholarGate. (2026, June 3). Predictive Site Location Modeling. ScholarGate. https://scholargate.app/en/archaeology/predictive-site-location
Which method?
Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.
- Space SyntaxArchaeology↔ compare
- Viewshed AnalysisArchaeology↔ compare