沈阳市细菌性痢疾疫情分类回归树分析

目的探讨细菌性痢疾疫情与气象因素的关系。方法以沈阳市1950-1996年细菌性痢疾月发病率为响应变量,以气温、气压、降水量、蒸发量等气象指标为预报变量,利用软件R2.31构建回归树模型,应用SPSS13.0进行相关分析。结果对发病率进行拟合的预测值与实际值相符合,优于传统线性回归方法。在温度指标体系中,最低温度指标与细菌性痢疾关系最为密切。多因素回归树中,最低气温、降水量和气压最显著。结论回归树模型是处理异质性数据的有效工具,既简单实用又便于解释,在疫情预测方面具有广阔应用前景。

Analysis on Bacillary Dysentery Incidence of Shenyang Based on CART

Objective To investigate the relationship between the incidence of bacillary dysentery and meteorological factors. Methods The incidence of bacillary dysentery from 1950 to 1996 in Shenyang was taken as response variable,and meteorological factors such as temperature,air pressure,precipitation and amount of evaporation as predictable variable. Software R 2.3.1 was used to establish the classification and regression tree(CART),and the correlation analysis was performed by using SPSS13.0. Results The model got very good fitting results. The predicted values accorded with the true values well and there was no statistical significance between them. Among all the temperature indexes,minimum temperature indexes were most related to bacillary dysentery incidence. Minimum temperature,precipitation and pressure were statistically influential factors in the multiple variable model. Conclusion CART is a powerful tool for dealing with data of heterogeneity; it is deem to have a good future in the prediction of infectious disease incidence for its simplicity and ease.