Kaldor J, Clayton D
Stat Med. 1985 Jul-Sep;4(3):327-35. doi: 10.1002/sim.4780040312.
Latent class analysis provides a useful framework for the analysis of epidemiological data which may have been mismeasured. In this paper, the latent class model is described in the context of logistic regression with categorical variables, and some examples of its application are provided. In particular, it is shown that adjustment for a misclassified confounding variable can be greatly improved by using the methods presented.
潜在类别分析为可能存在测量误差的流行病学数据的分析提供了一个有用的框架。在本文中，潜在类别模型是在具有分类变量的逻辑回归背景下进行描述的，并提供了一些其应用实例。特别是，结果表明，使用所提出的方法可以大大改进对误分类混杂变量的调整。
