主成分回归分析在细菌性痢疾与气象因素关系中的应用  

APPLICATION OF PRINCIPAL COMPONENT ANALYSIS AND MULTIVARIATE LINEAR REGRESSION IN THE ANALYSIS OF THE RELATIONSHIP BETWEEN INCIDENCE OF BACILLARY DYSENTERY AND METEOROLOGICAL FACTORS

作　　者：廖洪秀[1] 张强[1] 杜长慧 蒋小花[1] 邓长飞[1] 陈鑫[1] 

机构地区：[1]四川大学华西公共卫生学院,成都610041 [2]成都市疾病预防控制中心

出　　处：《现代预防医学》2009年第5期813-815,共3页Modern Preventive Medicine

摘　　要：[目的]研究细菌性痢疾发病与气象因素的关系;探讨主成分回归分析在细菌性痢疾发病与气象因素关系中的应用。[方法]将成都市1999～2005年细菌性痢疾发病情况进行描述性分析,其月发病数和同期气象因素资料进行相关分析、主成分回归分析。[结果]细菌性痢疾发病率有明显的季节性,病例以夏秋季(5到10月份)较多。相关分析表明细菌性痢疾月发病率与气温、降雨量呈正相关(r1=0.930,P1﹤0.05;r2=0.896,P2﹤0.05),与雾日呈负相关(r=-0.585,P﹤0.05);主成分回归分析建立了菌痢发病与气象因素的预测方程,对细菌性痢疾发病率影响较大的气象因素有风速、气温和降雨量。[结论]高温高湿易引起细菌性痢疾的高发;主成分回归建立的方程可对细菌性痢疾的月发病率进行预测。[Objective] To study the relationship between incidence of bacillary and meteorological factors, and investigate the application of principal component analysis and multiple linear regression in the analysis of relationship between incidence of bacillary dysentery and meteorological factors. [Methods] The incidence of dysentery in Chengdu was analyzed by descriptive study from 1999 to 2005. Data of monthly incidence of bacillary dysentery and monthly meteorological data during 1999 to 2005 were collected to establish a database for linear correlation, principal component and multiple linear regression analyses. [Results] Incidence of bacillary dysentery presented obvious seasonal fluctuation, and there were much more cases in summer and autumn （May to October） .There was a positive correlation between the incidence of dysentery and temperature and rainfall respectively （r1 = 0.930, P1﹤0.05; r2 = 0.896, P2﹤0.05）, while a negative correlation with fog days （r =-0.585, P﹤0.05）. Principal component analysis and multiple linear regression set up an estimated equation for the relationship between incidence of bacillary dysentery and meteorological factors, and showed some meteorological factors with greater influence on incidence of bacillary dysentery including wind speed, temperature, rainfall. [Conclusion] The higher incidence of bacillary dysentery is caused by higher air temperature and higher air humidity. An estimated equation based on principal component analysis and multiple linear regression could predict monthly incidence of bacillary dysentery.

关 键 词：细菌性痢疾 气象因素 主成分分析 

分 类 号：R195.1[医药卫生—卫生统计学]