Research On Spatiotemporal Analysis Of Hand, Foot And Mouth Disease In Wenzhou And Construction Of Early Warning Mode

Epidemiological descriptive statistics were first used to portray the pathological characteristics of HFMD in Wenzhou.And the geographic information management system of hand,foot and mouth disease in Wenzhou City from 2010-2018 was established,and on the established geographic information system,spatial geography trend surface analysis,kernel density analysis,kriging interpolation and spatial autocorrelation were used to comprehensively depict the spatial and temporal transformation pattern of hand,foot and mouth disease in Wenzhou City from 2010-2018.For the epidemiology analysis,if the year is used as the scale,the number of reported cases of HFMD in Wenzhou from 2010-2018 fluctuates around 2014.Using the month as a scale,the incidence of HFMD is strongly influenced by climate change,with April-July being the peak period for cases,with April together accounting for 57.6% of the annual cases and February being the lowest,accounting for only 1.6%of the annual cases.Analysis of the kernel density of healthcare facilities within Wenzhou showed a stepwise decline in healthcare sites from east to west.The kriging interpolation analysis of precipitation in Wenzhou City illustrates that precipitation in Wenzhou City has obvious seasonal characteristics.Trend surface analysis showed that the incidence of HFMD within Wenzhou had distinct spatial characteristics,with a higher number of incidences in the east than in the west,and a significant decline in HFMD from north to south.Spatial autocorrelation analysis showed that the distribution pattern of HFMD in Wenzhou from 2010-2018 showed aggregation;generalized G-statistical analysis showed that the type of aggregation near Lucheng District was high-value aggregation;local autocorrelation analysis showed that the hotspot area spread from south to central.Spearman correlation analysis showed that precipitation and temperature among meteorological factors,population density among socioeconomic factors,and The multiple linear regression model did not pass the p-value test and was discarded,and the predictive efficiency of the principal component regression model was low and should not be adopted.Finally,comparing the 14 machine learning algorithms established,the best prediction model is the Elman-GRNN model,with MAE of 3.141 and RMSE of 3.697.In general,the prediction energy efficiency of the combined prediction models is likely to be better than the respective single prediction models if the prediction ability of the antecedent models is not poor.