Distribution prediction and driving mechanism of Neolithic settlements in the Jing River Basin, Northwest China

The study of the human–land relationship during the Neolithic period helps to understand the mechanism of human–environment interaction. This study focuses on the Jing River Basin (JRB), the key node of the ancient Silk Road, and uses GIS technology and machine learning to build an archaeological prediction model (APM). The study aims to understand the migration, evolution, and influencing factors of human activities in the Neolithic period in the Guanzhong Basin. The results show that: (1) The XGBoost model demonstrated superior accuracy and stability in predicting settlement locations. (2) Altitude and distance from the large sites were key factors in the middle Neolithic period, while climatic deterioration heightened the importance of NDVI and roughness in the late Neolithic period. (3) A westward and southward migration of settlements from the middle to late Neolithic period, reflecting adaptive strategies to environmental change and potential cultural interactions with the Ganqing and Guanzhong regions.