Timo GeitlingerPlease use the format "First name initials family name" as in "Marie S. Curie, Niels H. D. Bohr, Albert Einstein, John R. R. Tolkien, Donna T. Strickland"
<p>Since its earliest application in the 1970s, network analysis has become increasingly popular in both theoretical and GIS-based archaeology. Yet, applications of material networks remained relatively restricted. This paper describes a specific kind of material network, so called similarity networks, and presents a new network analysis method to approach them: Similarity Network Fusion (SNF). </p>
<p>Most archaeological applications of material networks approach similarity by simply quantifying the number of co-occurrences of certain traits between two nodes, without considering the relative importance of these traits for the whole network. Statistical similarity measures have so far only been applied to a handful of case studies. The similarity network analysis outlined in this paper relies on SNF, a method common in genomic studies. SNF is based on the iterative integration of similarity networks derived from multiple datatypes for the same set of samples, allowing the application of a wide range of similarity indices. It has proven to be particularly robust to heterogenous and noisy datasets containing a small number of samples, but a large number of measurements, scale differences, and collection biases. </p>
<p>In a case study, I applied SNF to Early Bronze Age, Middle Bronze Age, and Late Iron Age burial sites in Dorset, resorting to data published by the Grave Goods Project. To enhance the understanding of the resulting networks and topological clusters, the network was spatially represented and the clusters were correlated with five further attributes; physiographic areas, the sex of the buried individuals, the ratio of objects per grave, whether isotopic analyses suggest that the buried people were local inhabitants or moved into the area of their burial, or whether there is an association between clusters and a finer chronological subdivision of sites. The network analysis and the topological clustering of the sites revealed at least two possible spatial clusters and two statistically significant correlations between clusters and further attributes of the burial sites. These results clearly suggest the great potential of SNF for analysing archaeological datasets, unveiling patterns within the archaeological record, and understanding the significance of these patterns for the structuration of the past landscape.</p>
Network Analysis, Similarity Networks, Dorset, Burial Goods, Prehistory, GIS, Similarity Network Fusion
