Scholarlink

<p>Brick and stone cultural relics (BSCR) are non-renewable heritage requiring scientific evaluation and proactive conservation. Conventional static approaches fail to reveal spatial patterns or predict distributions, hindering effective strategies amid urbanization. This study constructs a three-dimensional assessment framework (resource ontology, environmental stress, social value) with machine learning for 1448 BSCR in Hunan Province, China. Using kernel density, Geodetector, dominance-degree spatial prediction (DDSP), and dominance-degree zoning clustering (DDZC), results demonstrate: (1) Significant spatial clustering (NN ratio = 0.378), hotspots in Changsha-Zhuzhou-Xiangtan and Dongting Lake rim, 48.9% within 10 km of rivers; (2) Land-use type (<i>q</i> = 0.986) and population density (<i>q</i> = 0.957) are the strongest individual drivers, while land-use and aspect show the strongest interaction effect (power = 0.998); (3) DDSP achieves good predictive performance with AUC = 0.812 (95% CI: 0.794–0.829), though this indicates moderate predictive power, and should be interpreted with caution when identifying high-potential areas; (4) DDZC classifies four conservation priority levels.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Evaluation and prediction of dominance degree for brick and stone cultural relics resources

  • 期刊论文

  • 出版模式：

    开放获取

  • 发表日期：

    2026年4月22日

查看全文
  • Jingwei Hou,
  • Ji Zhou,
  • Yonghong He,
  • Bo Hou,
  • Tingliang Liu

摘要

Brick and stone cultural relics (BSCR) are non-renewable heritage requiring scientific evaluation and proactive conservation. Conventional static approaches fail to reveal spatial patterns or predict distributions, hindering effective strategies amid urbanization. This study constructs a three-dimensional assessment framework (resource ontology, environmental stress, social value) with machine learning for 1448 BSCR in Hunan Province, China. Using kernel density, Geodetector, dominance-degree spatial prediction (DDSP), and dominance-degree zoning clustering (DDZC), results demonstrate: (1) Significant spatial clustering (NN ratio = 0.378), hotspots in Changsha-Zhuzhou-Xiangtan and Dongting Lake rim, 48.9% within 10 km of rivers; (2) Land-use type (q = 0.986) and population density (q = 0.957) are the strongest individual drivers, while land-use and aspect show the strongest interaction effect (power = 0.998); (3) DDSP achieves good predictive performance with AUC = 0.812 (95% CI: 0.794–0.829), though this indicates moderate predictive power, and should be interpreted with caution when identifying high-potential areas; (4) DDZC classifies four conservation priority levels.