Multi‐isotope fingerprints in the bioapatite of archaeological skeletons are mostly superior over single isotope analyses for provenance studies. Gaussian mixture model (GMM) clustering is a novel tool for a similarity search among multidimensional data sets and at the same time permits the evaluation of the structural importance of particular isotopic ratios in the data set. We applied three GMM clustering experiments on multi‐isotope fingerprints—stable strontium (Sr), lead (Pb) and oxygen (O) isotopic ratios—established in 217 archaeological animal bones excavated along a specific transect across the European Alps. This reference region had been in use since prehistoric times by humans who crossed the Alps from north to south, and vice versa. The resulting clusters permit a spatial assignment of the specimens with a very high probability, in particular with regard to the geological complexity of the region. A combination of Sr with Pb stable isotopes led to an optimal differentiation between the southern and northern Alpine forelands that cannot be distinguished from each other by ⁸⁷Sr/⁸⁶Sr ratios alone, while the contribution of δ¹⁸O is not particularly high. The isotopic mapping and subsequent cluster analysis is suitable for the analysis of archaeological human finds and the reconstruction of the direction of transalpine mobility and trade.

中文翻译：

---

考古动物骨生物磷灰石中的多同位素指纹（O，Sr和Pb）：相似搜索以及在地质复杂的高山地区进行物源分析的适用性

在物源研究中，考古骨架生物磷灰石中的多同位素指纹图谱通常优于单一同位素分析。高斯混合模型（GMM）聚类是一种用于多维数据集之间进行相似性搜索的新颖工具，同时可以评估数据集中特定同位素比率的结构重要性。我们对多同位素指纹进行了三个GMM聚类实验-稳定的锶（Sr），铅（Pb）和氧（O）同位素比-在沿着欧洲阿尔卑斯山的特定断面挖掘的217个考古动物骨骼中建立。自从史前时代以来，这个参考区域就一直在使用，人们从北向南穿越阿尔卑斯山，反之亦然。由此产生的簇可以以很高的概率进行标本的空间分配，特别是在该地区的地质复杂性方面。锶与铅稳定的组合同位素导致了不能由彼此区分南北高山前陆之间的最佳分化⁸⁷ SR / ⁸⁶的Sr单独比率，而δ的贡献¹⁸ O为不是特别高。同位素测绘和随后的聚类分析适用于人类考古发现的分析以及跨高山迁移和贸易方向的重建。