Model‐based Cluster Analysis of Artefact Compositional Data

Cluster analysis is the most widely used multivariate technique in archaeometry, with the majority of applications being exploratory in nature. Model‐based methods of clustering have their advocates, but have seen little application to archaeometric data. The paper investigates two such methods. They have potential advantages over exploratory techniques, if successful. Mixture maximum‐likelihood worked well using low‐dimensional lead isotope data, but had problems coping with higher‐dimensional ceramic compositional data. For our most challenging example, classification maximum‐likelihood performed comparably with more standard methods, but we find no evidence to suggest that it should supplant these.     

Modelling Potential Socio-economic Determinants of Building Fires in South East Queensland

This paper investigates the relationships between areas of building fire incidence, levels of socio-economic disadvantage and the underlying socio-economic characteristics in the South East Queensland (SEQ) region, Australia. Disaggregated fire incident data was acquired from the Queensland Fire and Rescue Service (QFRS) and then aggregated to the Statistical Local Area (SLA) level. The Australian Bureau of Statistics (ABS) defined index of socio-economic disadvantage (called ‘SEIFA’) has been used as the basis to identify relationships between socio-economic disadvantage and building fires. A regression model was then developed to predict the incidence of building fires using a range of socio-economic variables. Five significant predictors were identified that include: i) percentage of unemployed, ii) proportion of Indigenous population, iii) families living in separate dwellings, iv) one parent, and v) parent families with children less than fifteen years of age. Results also show that the distribution of building fires varies markedly across the SEQ region, with some of the Brisbane inner suburbs, areas of high socio-economic disadvantage, and parts of inland SEQ associated with relatively high fire rates.     

An analytical study, using inductively coupled plasma (ICP) spectrometry, of samian and colour-coated wares from the Roman town at Colchester together with related continental samian wares

Over 200 sherds of samian and colour-coated pottery of late 2nd or early 3rd century date have been analysed by inductively coupled plasma (ICP) emission spectrometry. Most of this material was excavated at Colchester, England; the remainder is related East Gaulish material. Some clay samples were also analysed. The material divides, after multivariate statistical analysis, into 11 clusters. These accord well with known archaeological evidence and suggest assignments of origin where these are uncertain. In particular, the archaeological evidence for a link between Colchester and the Sinzig potters is supported.     

New Statistical Approaches to Intra‐individual Isotopic Analysis and Modelling of Birth Seasonality in Studies of Herd Animals

This paper introduces improved methods for statistically assessing birth seasonality and intra‐annual variation in δ¹⁸O from faunal tooth enamel. The first method estimates input parameters for use with a previously developed parametric approach by C. Tornero et al. The second method uses a non‐parametric clustering procedure to group individuals with similar time‐series data and estimate birth seasonality. This method was successful in analysing data from a modern sample with known season of birth, as well as two heterogeneous archaeological data sets. Modelling indicates that the non‐parametric approach estimates birth seasonality more successfully than the parametric method when less of the tooth row is preserved. The new approach offers a high level of statistical rigour and flexibility in dealing with the time‐series data produced through intra‐individual sampling in isotopic analysis.     

古陶瓷的成分测定,数据处理和考古解释

化学组成测定是古陶瓷科学研究的基础。作根据亲身实践,对几种常用的,或有发展前景的陶瓷化学组成的测量方法作了综评,定量地估计了它们的灵敏度,精密度和准确度。本还对古陶瓷化学组成数据多元统计处理中变量和样品的选择和舍弃,数据处理结果的表述,特别是作考古推论中的某些问题进行了探讨。     

Bulk element composition versus trace element composition in the reconstruction of an obsidian source system

It has become possible, through compositional analysis, to differentiate among discrete obsidian quarry and collecting localities within a single source area. The analytical potential of such information is considerable yet studies are seldom undertaken owing to the difficulties associated with trace element analysis.This study is intended to explore the feasibility of using bulk element data instead of trace element data in such compositional analyses. Techniques of bulk element analysis are far more accessible to archaeologists. The results of an experiment utilizing carefully controlled samples from the El Chayal obsidian source, Guatemala, are reported and discussed.     

Cluster analysis of mixed-level data: Pottery provenience as an example

Provenience or “sourcing” studies of archaeological objects most commonly employ quantitative measurements of amounts of chemical constituents in the artifacts and source materials. These are then used in statistical grouping procedures known as cluster analyses. Many additional attributes of artifacts and sources, yielding lower-level qualitative data, are also of potential importance in determinations of artifact origins. These attributes are usually neglected because of the longstanding difficulty in treating “mixed-level” data, that is, combining both qualitative observations and quantitative information in a single powerful statistical procedure. A series of cluster analyses were run on two data sets consisting of archaeological pottery, using mixed-level characterization data in the CLUSTAN program with Gower's coefficient of similarity. The clustering solutions are interpretable in terms of provenience and production strategies.     

On the similarity of tree‐ring patterns: Assessing the influence of semi‐synchronous growth changes on the Gleichläufigkeitskoeffizient for big tree‐ring data sets

The Gleichläufigkeitskoeffizient (GLK), or the percentage of parallel variation (%PV), is an often used non‐parametric similarity measure in dendrochronological research. However, when analysing big data sets using the GLK, this measure has some issues. The main problem is that it includes not only synchronous but also semi‐synchronous growth changes. These are years in which the growth in one of the compared series does not change in two subsequent years. This influences the GLK, often only slightly, but the larger the data set the stronger the effect. The similarity between tree‐ring series can be more objectively expressed by replacing the GLK with the synchronous (SGC) and semi‐synchronous growth changes (SSGC). The calculation is similar, since GLK = SGC + SSGC/2. Large values of the SSGC are indicative of possible anomalies or even errors. The SGC is much better suited than the GLK to describe similarity. The SGC should therefore be used to analyse big data sets, for clustering and/or dendroprovenance studies. It is recommended to combine the SGC with parametric measures.