Neural networks differentiate between Middle and Later Stone Age lithic assemblages in eastern Africa
Neural networks differentiate between Middle and Later Stone Age lithic assemblages in eastern Africa
复制标题
DOI：
10.1371/journal.pone.0237528
复制
发表时间：
2020-08-26
期刊：
影响因子：
3.7
通讯作者：
Blinkhorn, James
中科院分区：
文献类型：
--
作者：
Grove, Matt;Blinkhorn, James
石器时代中期到晚期的过渡标志着更新世晚期非洲人口生产和使用石器工具包的方式发生了重大变化，但在整个非洲大陆的不同方式，地点和时间都有体现。除了原材料使用模式的变化和人工制品尺寸的减小，人工制品类型的变化通常被用来区分中石器时代（MSA）和晚石器时代（LSA）的组合。目前的文件采用了定量分析框架的基础上使用神经网络来研究不断变化的星座之间的技术MSA和LSA组装从东非。训练网络集合以基于16种技术的存在或不存在来区分LSA组合与海洋同位素阶段3&4 MSA和海洋同位素阶段5 MSA组合。模拟被用来提取显着的指标和反指标技术，为每个集合类。经过训练的网络集合正确分类了94%以上的装配，并确定了7项关键技术，这些技术显著区分了装配类。这些结果澄清了时间变化的MSA和MSA和LSA组合在非洲东部之间的差异。
The Middle to Later Stone Age transition marks a major change in how Late Pleistocene African populations produced and used stone tool kits, but is manifest in various ways, places and times across the continent. Alongside changing patterns of raw material use and decreasing artefact sizes, changes in artefact types are commonly employed to differentiate Middle Stone Age (MSA) and Later Stone Age (LSA) assemblages. The current paper employs a quantitative analytical framework based upon the use of neural networks to examine changing constellations of technologies between MSA and LSA assemblages from eastern Africa. Network ensembles were trained to differentiate LSA assemblages from Marine Isotope Stage 3&4 MSA and Marine Isotope Stage 5 MSA assemblages based upon the presence or absence of 16 technologies. Simulations were used to extract significant indicator and contra-indicator technologies for each assemblage class. The trained network ensembles classified over 94% of assemblages correctly, and identified 7 key technologies that significantly distinguish between assemblage classes. These results clarify both temporal changes within the MSA and differences between MSA and LSA assemblages in eastern Africa.