The Use of Simulation-based inference methods in the Parameter Estimation of Factors Governing the Location of Known Ancient Roads

基于仿真的推理方法在已知古道位置影响因素参数估计中的应用

• 批准号: 2441218
• 金额: --
• 依托单位: University of Cambridge
• 依托单位国家: 英国
• 项目类别: Studentship
• 财政年份: 2020
• 资助国家: 英国
• 起止时间: 2020 至 无数据
• 项目状态: 已结题
• 来源: https://gtr.ukri.org/projects?ref=studentship-2441218
• 关键词: Use; Simulation; based; inference; methods

By studying past movements, we can understand the flow of people, ideas, objects, and information from place to place.In particular, roads provide important information about the structure of a society and how societies inhabited thelandscape. In the last two decades, the use of spatial modelling and Geographic Information Systems to reconstructancient roads and identify factors governing the construction of known roads has increased In particular, Least Cost Path(LCP) analysis, which finds the optimal path between two geographic points (Hardin et al., 2012) has becomewidespread in its application. There is currently considerable debate over the best methodological application of LCP inarchaeology. Though Rahn (2005) states that LCP analysis represents a well-established methodology, Herzog (2013,184) considers this by "no means trivial" and highlights that it is still necessary to investigate the parameters used withinthe LCP calculation. However, as models become more complex, the impact of each cost factor on the computed LCPbecomes more difficult to assess. To overcome the limitations in the current approach to LCP analysis and to enhanceour understanding of the factors dictating the emergence and evolution of roads, this research proposes simulation based inference methods, through the use of Bayesian statistics. The Bayesian approach to modelling will allow for theestimation of factors that plausibly governed where ancient roads were located.

通过研究过去的运动，我们可以了解人、思想、物品和信息从一个地方到另一个地方的流动。特别是，道路提供了关于社会结构和社会如何居住在景观中的重要信息。在过去的二十年中，使用空间建模和地理信息系统来重建旧道路并识别控制已知道路的建设的因素已经增加。特别是，最小成本路径（LCP）分析，其找到两个地理点之间的最佳路径（哈丁等人，2012年）已在其应用中得到广泛应用。目前有相当大的辩论最好的方法应用LCP在考古学。尽管Rahn（2005）指出LCP分析代表了一种成熟的方法，但Herzog（2013，184）认为这“绝非微不足道”，并强调仍有必要研究LCP计算中使用的参数。然而，随着模型变得越来越复杂，每个成本因素对计算出的LCP的影响变得越来越难以评估。为了克服在目前的LCP分析方法的局限性，并提高我们的理解的因素支配的道路的出现和演变，本研究提出了基于模拟的推理方法，通过使用贝叶斯统计。贝叶斯建模方法将允许估计可能决定古代道路所在位置的因素。