【作者】 谭慭莘；

【导师】 田考聪；

【作者基本信息】 重庆医科大学 ， 流行病与卫生统计学， 2005， 硕士

【摘要】 目的 针对库区某县三种介水传染病发病情况及影响因素研究,探讨合理的预测模型,为制定库区某县介水传染病的预防和控制措施提供科学依据,同时为三峡库区人群三种介水传染病发病情况预测模型研究奠定基础。 方法 运用线性回归、灰色关联分析、灰色预测和时间序列分析方法(ARIMA 乘积模型,简约的季节周期模型)分别拟合分月疾病发病预测模型、分季度疾病发病预测模型和分监测点疾病发病预测模型,对该县 2000 年至 2004 年三种介水传染病(甲型肝炎、细菌性痢疾和其它感染性腹泻)发病情况和影响因素进行分析。 结果 1、本次研究发现 2000 年至 2004 年甲型肝炎、细菌性痢疾和其它感染性腹泻发病率呈周期波动分布,每年的夏、秋季节发病率高于冬、春季节的发病率;从 2000 年至 2004 年三种介水传染病的发病呈逐年上升趋势;2003 年 5 月该县发生甲型肝炎暴发流行,该月的发病率高于其它年的同期水平和该年其它月的发病水平。 2、 线性相关分析和灰色关联分析两种方法发现:硝酸盐氮、总氮和生化需氧量为水质监测指标中影响该县 2000 年至 2004 年三种介水传染病发病的主要因素。 3、预测三种介水传染病月发病率,本研究建立了分月三种介水传染病发病情况的 ARIMA(1,0,0) x(0,1,0)12乘积模型,其模型的精度和预测效果较为理想; 4、预测三种介水传染病季度发病率,本研究建立了预测分季度三更多还原

【Abstract】 Objective Aim at the incidence of three kinds of water-bome infection and these influence factors, the more reasonable predictive model will be investigated and established, which will offer scientific foundation to establish prevention and controlling measurement. Methods Applying linear regress analysis, gray relevancy analysis, gray forecasting and time series analysis (including: ARIMA product model, simple seasonal periods model), the incidence of three kinds of water-bome infection (including: Hepatitis A, Bacillary Dysentery, others Infectious Diarrhea) and influencing factor of these disease were be studied from 2000 to 2004 year. Results 1 、 The incidence of diseases (including: Hepatitis A, bacillary dysentery, others infectious diarrhea) is seasonally fluctuant-distributing form 2000 to 2004 year. In summer and in autumn, the incidence of the diseases is higher than the incidence of the diseases in spring and winter. The hepatitis A broke out on May 2003. 2、Nitrate nitrogen, total nitrogen and BOD are three main factors of the three kinds of water-bome infection in the area ,according to the results of linear regress analysis and gray relevancy analysis; 3、The results of the ARIMA(1,0,0) x(0,1,0)12 model is good.The models may reasonably predict in future; 4、The gray GM (1,1) model and the simple seasonal periods model are good. Comparing the gray GM (1,1) model with the simple seasonal periods model, the simple seasonal periods model’s results is better than GM (1,1) model’s, but GM (1,1) model may reasonably predict in future; 5、Comparing the multipul stepwise regress analysis with the GM(0,3) model, the GM(0,3) model’s MAPE is less than the multipul stepwise regress. The GM (0,3) model may reasonably predict in future. Conclusions 1、The models of ARIMA and the simple seasonal periods model can be used to forecast for these diseases incidence with high prediction of short-term series. 2、Applying the models of the GM(0,3) ,the data of the water quality surveillance can forecast for the incidence of the diseases with high precision .更多还原