基于遥感技术和时间序列分析法预测皖北疟疾的发病率

发布时间：2018-09-01 09:46

【摘要】：目的探讨皖北疟疾的发病率与地表温度(land surface temperature,LST)、归一化植被指数(normalized difference vegetation index,NDVI)的关联性,评价用LST、NDVI对疟疾发病率自回归移动平均模型(autoregressive integrated moving average model,ARIMA)预测结果进行校正的效果。方法以皖北五县为研究现场,收集各县2004-2011年的疟疾疫情数据及LST、NDVI等遥感图像资料,提取、合成遥感相关指标;运用SPSS 17.0软件进行统计学处理。结果 ARIMA模型对2010年各月份的预测结果较报告发病率高(平均误差=0.721/10万)。多因素分析结果显示,当地的疟疾发病率与近三个月的平均LST(lst_(_012),β=0.295)及之前两个月的平均NDVI(ndvi_(_12),β=0.280)有关联(P0.001);将二者作为校正因子(相对贡献为2∶1时)对2010年的预测结果进行校正,平均误差缩小为0.018/10万。以2004-2010年的发病率数据再次拟合并筛选ARIMA模型,并以2011年的疟疾报告发病数据为参照,再次评价lst_(_012)与ndvi_(_12)对模型预测结果的校正效果;发现校正后的预测误差(0.001/10万)低于校正前的误差(0.293/10万)。结论 ARIMA模型能较好地用于该地疟疾发病率的拟合与预测,环境遥感替代指标LST、NDVI可在一定程度上改善ARIMA模型的预测效果。
[Abstract]:Objective to investigate the correlation between malaria incidence and surface temperature (land surface temperature,LST) and normalized vegetation index (normalized difference vegetation index,NDVI) in Northern Anhui Province, and to evaluate the effect of LST,NDVI in correcting the predicted results of (autoregressive integrated moving average model,ARIMA (an autoregressive moving average model) of malaria incidence. Methods the data of malaria epidemic situation and remote sensing image such as LST,NDVI were collected from 2004 to 2011 in five counties of northern Anhui, and the related indexes of remote sensing were extracted and synthesized, and the statistical data were processed by SPSS 17.0 software. Results the predicted results of ARIMA model for each month of 2010 were higher than the reported incidence (average error = 0.721 / 100 000). Multivariate analysis showed that the local incidence of malaria was associated with the average LST (lst_ (lst_), 尾 -0.295) in the last three months and the average NDVI (ndvi_ (12), 尾 0.280) in the previous two months (P0.001), which were used as correction factors (at 2:1) to correct the predicted results for 2010. The average error shrank to 0.018 / 100, 000. Based on the incidence data of 2004-2010, the ARIMA model was selected and compared with the malaria incidence data reported in 2011. The correction effect of lst_ (012) and ndvi_ (12) on the predicted results of the model was evaluated again. The corrected prediction error (0.001 / 100,000) is lower than the pre-corrected error (0.293 / 100,000). Conclusion ARIMA model can be used to fit and predict the incidence of malaria in this area, and the environmental remote sensing replacement index LST,NDVI can improve the prediction effect of ARIMA model to some extent.
【作者单位】：安徽中医药大学公共卫生与全科医学教研室;安徽省疾病预防控制中心地方病与寄生虫病防治科;
【基金】：国家自然科学基金(30901255) 安徽省高校优秀青年人才支持计划项目(2014-077)
【分类号】：R531.3;R181.3

【相似文献】