A settlement analysis on Dutch inhabitants throughout the region of the southern-Netherlands  during the period of 3000-800 BC, assisted by a maximum entropy species distribution  model.

This thesis will study the predictive modelling technique of Maximum Entropy modelling and see it integrated in a settling study of the prehistoric southern Netherlands. Maximum entropy is a machine learning algorithm, specifically a log-linear model which operates under an Ockham’s razor approach where it determines probability and the associated probability distribution based on the least amount of assumptions required while still adhering to the set input parameters. It is thus ideally suited for predictive modelling goals as no bias can occur, only when manually inputted. While normally used in ecological habitat studies to study the spread of certain animal or plant species, it has been used in recent times to study settlement patterns in the field of archaeology. This Maximum Entropy model will attempt to showcase possible settlement locations for Bronze Age peoples living in the southern Netherlands. A brief overview of the history of the Bronze Age southern Netherlands will...Show moreThis thesis will study the predictive modelling technique of Maximum Entropy modelling and see it integrated in a settling study of the prehistoric southern Netherlands. Maximum entropy is a machine learning algorithm, specifically a log-linear model which operates under an Ockham’s razor approach where it determines probability and the associated probability distribution based on the least amount of assumptions required while still adhering to the set input parameters. It is thus ideally suited for predictive modelling goals as no bias can occur, only when manually inputted. While normally used in ecological habitat studies to study the spread of certain animal or plant species, it has been used in recent times to study settlement patterns in the field of archaeology. This Maximum Entropy model will attempt to showcase possible settlement locations for Bronze Age peoples living in the southern Netherlands. A brief overview of the history of the Bronze Age southern Netherlands will be given, where natural, economic and cultural factors will be highlighted for settlement patterns or trends that occur during research. These factors range from geographical information of fertile areas or factors such as the proximity to commonly traded items or cultural information like a late Neolithic/Bronze Age phenomenon of cultural deposition where certain items, usually luxury items, are deposited in certain types of landcover such as peat soils for religious/ritualistic end goals. These mentioned factors will be analysed and used to make a predictive model of the entirety of the southern Netherlands where a distinction will be made between natural, economic and cultural factors and how these interact or complement one another in this research. This data will be analysed and fitted in a visual display which will rank certain areas from 0 to 1, 1 being an ideal settlement location and 0 not suited as a location to settle for Bronze Age peoples residing in the southern-Netherlands. Ideal settlement locations are classified as ideal when the combination of environmental parameters are suited for settlement purposes. For example, a combination of proximity to water on elevated areas close to cultural or economic conditions will be classified as an ideal location for settlement. Not suited locations are classified when there is no ideal combination of environmental parameters to be found. These rankings are based on permanent settlement, mainly focusing on agricultural subsistence strategies instead of subsistence strategies as hunting/foraging as these settlements tend to be temporal rather than permanent. Post analysis and fitting of the model, a comparative part to this research to analyse the fit of the maximum entropy model compared to models not created through machine learning methodology like the IKAW (Indicatieve Kaart Archeologische Waarden), a predictive model of the Netherlands indicating the find locations for archaeological objects on a bigger scale (1:50 000) will be integrated. While the IKAW is an indicative map on a scale of 1 : 50 000, this maximum entropy predictive model is 1:10 000 scale making it more ‘accurate’ than the IKAW, yet still useful for comparative research.Show less

Leiden University Student Repository

Documents

In Collections

A settlement analysis on Dutch inhabitants throughout the region of the southern-Netherlands during the period of 3000-800 BC, assisted by a maximum entropy species distribution model.