简体繁体 English

如何在 R 中使用分类变量和连续变量执行潜在配置文件分析？

[英]How do I perform a Latent Profile Analysis with both categorical and continuous variables in R?

原文 2020-06-04 16:49:27 4 1 r

I have 4 continuous variables (masculinity, femininity, partner's masculinity, partner's femininity) and 2 categorical variables (gender identity, partner's gender identity) and I'd like to perform an LPA.我有 4 个连续变量（男性气质、女性气质、伴侣的男性气质、伴侣的女性气质）和 2 个分类变量（性别认同、伴侣的性别认同），我想执行 LPA。 In my readings, it seems possible to use both categorical and continuous indicators for an LPA, but I haven't found a resource for actually doing it.在我的阅读中，LPA 似乎可以同时使用分类指标和连续指标，但我还没有找到实际使用的资源。 I have attempted using the package tidyLPA but it doesn't allow me to specify which variables are categorical.我曾尝试使用 package tidyLPA，但它不允许我指定哪些变量是分类的。

Am I correct to assume LPA can be used for both categorical and continuous indicators?我是否正确假设 LPA 可用于分类和连续指标？
Is there a package in R that can handle this analysis? R 中是否有可以处理此分析的 package ？

Thank you!谢谢！

1 个解决方案

Am I correct to assume LPA can be used for both categorical and continuous indicators?我是否正确假设 LPA 可用于分类和连续指标？

From the documentation of the package...从 package 的文档中......

Note that LPA works best with continuous variables (and, in some cases, ordinal variables), but is not appropriate for dichotomous (binary) variables.请注意，LPA 最适用于连续变量（在某些情况下，还有序数变量），但不适用于二分（二元）变量。

Since categorical in this sense is much closer to binary than ordinal (unless you failed to mention that you meant ordinal not categorical) odds are strong it's inappropriate.由于在这个意义上分类比序数更接近二元（除非你没有提到你的意思是序数不是分类）可能性很大，这是不合适的。

Is there a package in R that can handle this analysis? R 中是否有可以处理此分析的 package ？

See the comment from Peter above but it appears tidyLPA just wraps mclust and that mclust is capable of different models including categorical variables see for example the EDDA example.请参阅上面 Peter 的评论，但似乎tidyLPA只是包装mclust ，并且mclust能够处理不同的模型，包括分类变量，例如EDDA示例。