Probabilistic logic analysis of the highly heterogeneous spatiotemporal HFRS incidence distribution in Heilongjiang province (China) during 2005-2013

摘要

Author summary Heilongjiang Province reported the largest number of HFRS cases in China. Previous studies focused on identifying the severe HFRS outbreak regions, exploring the relative impact of environmental factors, forecasting HFRS cases etc. However, the study of the spatiotemporal spread dynamics and patterns of HFRS is still lacking, which is the focus of the present study. This study proposed a novel mapping technique (i.e., class-dependent Bayesian Maximum Entropy, cd-BME) for studying the distribution of HFRS, overcoming the highly heterogeneous features of HFRS data; and, probabilistic logic notions (stochastic indicators) were employed to study the spatiotemporal dependency of HFRS incidence and draw conclusions regarding the HFRS spread under conditions of uncertainty. By dividing the original HFRS data into four classes in terms of percentiles, the cd-BME exhibited better performance in mapping HFRS distribution than the standard (class-independent) BME technique and the mainstream inverse distance technique. Regarding the maps of HFRS distribution, the point outbreak character dominated the HFRS spread at high incidence levels, whereas the lowest incidence level covered more than 2/3 of Heilongjiang Province. Certain HFRS incidence generally occurred between intraclass or neighbor classes. The probabilities of HFRS transition between incidence classes with various spatial distances and temporal instants can be found in the HFRS indicator plots. The above comprehensive information can allow a better understanding of the spatiotemporal HFRS spread mechanisms and further improve HFRS decision-making, management and control.