We propose a root-Gaussian Cox process for the analysis of data that is aggregated in space and where the boundaries of geographic regions change over time. A generalized EMS algorithm exploits sparse structure resulting in significant computational gains. Inference for a spatial intensity surface is conducted using a non-central chi-square distribution. A simulation study over the state of Kentucky demonstrates the efficacy of the methodology. An application to a large region in India allows the investigation of the spatial structure of the incidence of Plasmodium falciparum, a severe strain of malaria.

中文翻译：

---

根高斯 Cox 过程和广义 EMS 算法

我们提出了一个根高斯 Cox 过程，用于分析在空间中聚合的数据以及地理区域边界随时间变化的数据。广义 EMS 算法利用稀疏结构，从而获得显着的计算收益。空间强度表面的推断是使用非中心卡方分布进行的。对肯塔基州的模拟研究证明了该方法的有效性。应用到印度的一个大区域，可以调查恶性疟原虫（一种严重的疟疾毒株）发病率的空间结构。