基于GA-BP神经网络模型的登革热时空扩散模拟

发布时间：2018-04-02 06:25

本文选题：登革热　切入点：时空数据挖掘　出处：《中国图象图形学报》2015年07期

【摘要】：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。
[Abstract]:Objective Dengue fever is a global public health problem. It is a new method to effectively prevent and control dengue fever by exploring the spatio-temporal characteristics of dengue fever from the perspective of spatio-temporal data analysis and constructing a model of spatio-temporal process of dengue fever. Methods based on spatio-temporal data mining, spatio-temporal process modeling, integrated environment, meteorology, geography and population, the spatial correlation of dengue fever and the spatial autocorrelation of dengue cases were analyzed, and the influencing factors of dengue fever were excavated. The genetic algorithm (GA) is introduced to improve the BP neural network model for spatiotemporal simulation of dengue fever. The results show that the spatio-temporal diffusion of dengue fever is related to temperature, humidity, residential land, traffic, etc. There was a significant correlation between population density, significant autocorrelation among dengue cases, nonlinear relationship between occurrence, diffusion and environment, meteorology, geography and population. The improved GA-BP neural network model was used to simulate the spatial and temporal diffusion of dengue fever, and the root mean square error was 0.081. Conclusion Dengue fever occurs and diffusion is the result of multiple factors. The GA-BP neural network model can effectively simulate the spatio-temporal process of dengue fever. This model can also be used to simulate other Aedes mosquito infectious diseases.
【作者单位】：华南师范大学地理科学学院;
【分类号】：R512.8;TP18

【参考文献】