LAMPS and Department of Mathematics and Statistics, York University, Toronto, ON, Canada.
J Med Entomol. 2019 Jan 8;56(1):65-71. doi: 10.1093/jme/tjy118.
Mosquito trap counts are heavily influenced by environmental factors such as temperature and precipitation. However, some important geographic factors, such as land use and elevation of a particular site, are often either not recorded or simplify not observable. This is a major issue in building a predictive model for the mosquito trap counts over time across a particular region. The collective impact of all unobservable factors for one particular site is estimated by a hidden dimension method. Application to mosquito trap counts in Peel Region has shown that our model can significantly improve the modeling accuracy of the generalized linear model. This method may provide a significantly better characterization of the spatiotemporal distribution of mosquito (Diptera: Culicidae) abundance in areas with green lands or open spaces.
蚊虫诱捕器计数受环境因素（如温度和降水）的影响较大。然而，一些重要的地理因素，如特定地点的土地利用和海拔高度，通常要么未被记录，要么难以观察。这是在特定区域内随时间构建蚊虫诱捕器计数预测模型的一个主要问题。通过隐藏维度方法来估计特定地点所有不可观测因素的综合影响。在皮尔地区蚊虫诱捕器计数中的应用表明，我们的模型可以显著提高广义线性模型的建模精度。该方法可能会更好地描述绿地或开阔空间地区蚊虫（双翅目：蚊科）丰度的时空分布。
