论文部分内容阅读
目的登革热是一个全球性公共卫生问题,从地理学时空数据分析的视角,探究登革热的时空特质、构建登革热时空过程模型,是有效预防、控制登革热的新方法、研究新热点。方法基于时空数据挖掘、时空过程建模,综合环境、气象、地理、人口4大因素,分析登革热的空间相关性及登革热病例的空间自相关,挖掘登革热影响因子;针对BP(back propagation)神经网络模型易陷入局部最优的缺陷,引入遗传算法(GA)改进BP神经网络模型,用于登革热时空模拟。结果登革热的时空扩散与温度、湿度、居民地、交通、人口密度呈显著相关;登革热病例之间呈显著自相关;登革热发生、扩散与环境、气象、地理、人口中的多种因子存在非线性关系;利用改进的GA-BP神经网络模型模拟登革热时空扩散,均方根误差达到0.081。结论登革热发生、扩散是由多种因素综合影响的结果;GA-BP神经网络模型能够有效模拟登革热时空过程;此模型同样适用于其他伊蚊类传染病的模拟。
Objective Dengue is a global public health issue. From the perspective of geo-spatial-temporal data analysis, it explores the temporal and spatial characteristics of dengue and builds a spatio-temporal process model of dengue, which is a new method to effectively prevent and control dengue fever. Methods Spatial-temporal data mining, spatio-temporal process modeling, comprehensive environmental, meteorological, geographical and population factors were used to analyze the spatial correlation of dengue fever and the spatial autocorrelation of dengue fever cases and to explore the influencing factors of dengue fever. For BP (back propagation) neural network The model is easy to fall into the local optimal defect. Genetic algorithm (GA) is introduced to improve the BP neural network model for the space-time simulation of dengue fever. Results The temporal and spatial distribution of dengue fever was significantly correlated with temperature, humidity, residential area, traffic and population density. There was significant autocorrelation between dengue fever cases. There was a non-linearity between dengue fever occurrence and spread and many factors in environment, meteorology, geography and population The GA-BP neural network model was used to simulate the spatiotemporal diffusion of dengue, and the root mean square error was 0.081. Conclusion The occurrence and spread of dengue fever are the result of a combination of factors. The GA-BP neural network model can effectively simulate the spatio-temporal process of dengue. This model is also suitable for the simulation of other mosquito-borne infectious diseases.