A Multivariate Nonlinear Regression Analysis Of Schistosomiasis In Nanchang County Based On The Administration Of Village Scale

Using the classical statistics and GIS/RS spatial analysis software to make an analysis on the difference of the factors which influence the schistosomiasis within Nanchang Country,providing techincal method for schistosomiasis control.1.Collecting The data of schistosomiasis epidemic situation of Nanchang Country in2009.2.Download the TM remote sensing image of Poyang Lake inJune,2009and the GIS thematic data of Nanchang County.1. Investigated83administrative village of water contacting which include seasons, days, times,numbers and location in Nanchang County through the questionnaire.2. Extractingriver system and the boundary line data of Poyang Lake and vector layers with Erdas9.3software to build geographic information system database.3. Statistical annlysis. Using Excel to date preprocessing of Water contact. Using SPSS17.0to make a related analysis on the schistosomiasis epidemic factors.then establish a regression model between five related factors and infections of eachvillage. Exploration of the influence of Nanchang County village administrationunit of schistosomiasis epidemic factors differences and inherent mechanism.Results:1.Calculate the infected water contacting index of each village through the questionnaire survey data.Using water contact index and schistosomiasisby regression analysis obtain that the three function models the most fitting which R value is0.86,can explain86%of the total variance.The result is very good.2.Using Erdas and Arcgis to acquire the distance between administrative village and the Poyang lake,water area of each village and density of administrative village’s boundary.3. Through the five factors of schistosomiasis infection and the number of correlation analysis, in which the water contact index and schistosomiasis infections of the simple correlation coefficient maximum(achieve0. 877) followed by administrative village water area(the simple correlation coeffi cient achieve0.671).4.Set a multivariate nonlinear regression model by SPSS Quadratic model is: Y=61.111+1.16E-07X1X4-0.005X4+3.98E-6X1X5Cubic model is: Y=28.489+3.35E-10X1X4X5+2.07E-06X1X5Biquadratic model is: Y=30.88+1.50E-12X25X1X4+8.65E-12X43X2-1.30E-18X14-3.04E-16X42X3X5+2.01E-13X13where X1means the infected water contacting index; X2means density of administrati ve village’s boundary; X3means water area of each administrate village; X4means th e distance between administrative village and the Poyang lake; X5means the numbe r of infected cattles.The R2of quadratic model, cubic model and biquadratic model is0.85,0.872and0.884,which means a good fitting effect. The Biquadratic model can expl ain88.4%of the total variance, which is the best model of all.5. According to the r eference information,setting a weight for each variable of the infected water cont acting index.Quantifying the contacting infected water index by assignment metho d can calculate the cotacting infect water index quickly and accurately.6. Single facto r related analysis shows that:the cotacting infected water index, the number of infecte d cattle,water area of each village and density of administrative village’s boundary ha ve a strong positive linear correlation with the number of infected people.The distanc e between administrative village and the Poyang lake is not significant in linear correl ation with the number of infected people.7.The study shows that with the coaction o f multiple factors,the infection of NaChangXian administrative villages is produce d.