论文部分内容阅读
目的 探讨树型分类方法在流行病学队列资料分析中的应用价值。方法 依据人群的同质性将总人群根据某一变量 (在流行病学分析中 ,通常为危险因素 )分成不同的亚人群 (在不同的亚人群中疾病的发生概率不同 )。结果 各危险因素在疾病发生中的作用、作用方式 ;危险因素间的交互作用的估计 ;疾病高危人群的筛选。结论 树型分类对多元分类资料进行分析时 ,不涉及参数推断 ,原理简单 ,尤其适用于有共线性的资料处理。在疾病的危险因素分析、以及变量筛选等方面将具有广泛的应用前景。
Objective To explore the application value of tree classification method in epidemiological cohort data analysis. Methods According to the homogeneity of the population, the total population is divided into different subpopulations according to a variable (usually a risk factor in epidemiological analysis) (the probability of occurrence of the disease is different in different subpopulations). RESULTS: The role of various risk factors in the development of the disease, the mode of action, the estimation of the interactions between the risk factors, and the screening of high-risk populations. Conclusion When tree classification is used to analyze multivariate classification data, parameter inference is not involved and the principle is simple. It is especially suitable for data processing with collinearity. It will have broad application prospects in the analysis of disease risk factors and variable screening.