一篇德国人在Journal of Clinical Medicine （临床医学）上发的文章，关于GLMM中R2的描述，我觉得非常到位，希望用这个R2能参考一下：
"2.3.3. Variance Explanation and Hierarchical Partitioning Derived from Adjusted Analysis
The pseudo-R2 method suggested by Nakagawa and Schielzeth [41] was used to estimate the effect size using the variance explained by LMMs. The explained variance was calculated as R2 using the r.squaredGLMM function in the MuMIn package [42]. The marginal R2 (mR2) represents the proportion of the total variance explained by fixed effects only. In contrast, the conditional R2 (cR2) estimates the variance the entire model explains (both fixed and random effects). Hierarchical partitioning was performed using the glmm.hp package to decompose variance [Lai et al., 2022; Lai et al., 2023]."
https://www.mdpi.com/2077-0383/13/16/4754
转载本文请联系原作者获取授权，同时请注明本文来自赖江山科学网博客。
链接地址：https://wap.sciencenet.cn/blog-267448-1446784.html?mobile=1
收藏
