Spatial modeling of Dengue prevalence and kriging prediction of Dengue outbreak in Khyber Pakhtunkhwa (Pakistan) using presence only data
Spatial modeling of Dengue prevalence and kriging prediction of Dengue outbreak in Khyber Pakhtunkhwa (Pakistan) using presence only data
复制标题
利用现场数据对巴基斯坦开伯尔-普赫图赫瓦登革热流行的空间建模和登革热暴发的克里格法预测
DOI：
10.1007/s00477-020-01818-9
复制
发表时间：
2020-05-31
影响因子：
4.2
通讯作者：
Shakir, Muhammad
中科院分区：
文献类型：
--
作者：
Ahmad, Hammad;Ali, Asad;Shakir, Muhammad
在2017年8月至10月期间，巴基斯坦开伯尔-普赫图赫瓦省发生了登革热大爆发。所有主要城市和农村地区都报告了病例，但白沙瓦受到的打击更为严重，一半以上的病例属于白沙瓦市中心。流行模式显示，登革热病例主要报告在平原地区，也有低海拔山区。我们采用最大熵原理来建立登革热存在和背景数据的基本分布。通过建立登革热风险的空间结构模型并在考虑到一些最重要的协变量的情况下估计具有相应不确定性的预测图，进行了地统计分析。使用二项克里金法和二进制逻辑漂移创建预测图。对全省和分区域进行了分析，以便更仔细地了解地方一级的空间分布情况。我们的结果表明，我们的方法表现良好。病媒分布，人口密度，距离道路被发现显着影响风险的空间分布，并提供了非常翔实的模式。
During the span of August-October, 2017 a major outbreak of Dengue fever happened in Khyber Pakhtunkhwa province of Pakistan. Cases were reported from all the major cities and rural areas, but Peshawar was more severely hit with more than half of the total cases belonging to central Peshawar city. The epidemic patterns reveal that dengue fever cases were mostly reported for plain areas and also low altitude mountainous regions. We employed the principle of maximum entropy to establish the underlying distribution of dengue presences and background data. A geostatistical analysis was conducted by modelling the spatial structure of the dengue fever risk and estimating the prediction maps with corresponding uncertainty taking into account some of the most significant covariates. The prediction maps were created using binomial kriging with a binary logistic drift. The analysis was carried out for the whole province as well as subregions to have a closer look of the spatial distribution at local level. Our results show that our methodology performed well. Vector distribution, population density, and distance to roads were found to significantly affecting the spatial distribution of risk and gives very informative pattern.