Macro Factor Analysis And Prediction Of Tuberculosis Incidence In China Based On Improved GTWR Model

Pulmonary tuberculosis,as an important disease affecting the health of Chinese people,has spatial and temporal heterogeneity.At present,our country in the outbreak era,various social macro factors have undergone corresponding changes.In order to achieve tuberculosis by 2035,so it is necessary to understand the characteristics of tuberculosis and different macro factors on the incidence of tuberculosis in China time and space characteristics,accurately predict the incidence trend,develop favorable prevention and control measures,so as to provide the basis for eliminate tuberculosis.This thesis mainly solves the following problems:(1)the spatial and temporal heterogeneity of the influence of different macro factors on tuberculosis incidence in 31 provinces and cities in China is still unclear;(2)only time factors and GTWR model can only be fit within sample without extrapolated,making is difficult to accurately predict the future incidence trend.For question 1,this thesis first introduces the basic spatial and temporal characteristics of pulmonary tuberculosis incidence in China from 2009 to 2018,Through the factors that affect pulmonary tuberculosis,using the non-parametric Spearman rank correlation test,21 indicators of 6 categories were selected as the macro factor dataset,Principal component dimensionality reduction and multiple collinearity test.Then,based on the GTWR model,Fitting the relationship between pulmonary TB incidence and macroscopic factors,estimating the coefficient function,thus revealing the spatiotemporal heterogeneity,compared with the ordinary linear regression(OLR)model and the geographic weighted regression(GWR)model.Finally,the results are estimated based on the coefficient function of the model to explore the influence of different macro factors on different regions.The results show that:(1)the fitting effect of GTWR model is better than that of GWR model and OLR model;(2)the influence of various macro factors on tuberculosis is cultural literacy,economy,environment,health,transportation,urban and rural differences;(3)there are significant spatial and temporal heterogeneity on the incidence of tuberculos is in different provinces and cities in China.In view of this,the provinces and cities should be according to the characteristics of various factors,according to local conditions to formul ate corresponding measures.In view of problem 2,this thesis considers the influence of six macro factors and other uncertain information,improves the GTWR model,and proposes a new statistical method,namely the GTWR-Markov-GM(1,1)model.With the help of GTWR model,the model added six macro factors affecting the incidence of pulmonary tuberculosis,eliminated the random fluctuations caused by uncertain information in the GTWR model through Markov model,and modified the incidence fitting value of GTWR model.On this basis,GM(1,1)model was established for 31 provinces and cities.After inspection,the accuracy level of the new model is qualified and can be used for extrapolation prediction.Through external verification of the incidence data from 2017 to 2018,the new model predicted better than the single GM(1,1)model.In this thesis,using the GTWR-Markov-GM(1,1)model and using the data from 2009-2018 to predict the incidence of pulmonary tuberculosis in 31 regions,we found that the incidence in Guizhou,Qinghai,Tibet and Xinjiang was increasing,while other provinces and cities were decreasing.The results of this thesis can provide a reference for establishing the prediction and early warning mechanism of pulmonary tuberculosis pathogenesis,and the model can also be extended to the study of other infectious diseases.