Mixed-Effects Multinomial Logistic Regression

Shuyuan Liu

2016 7 25

0 个回答

8 次查看（30 天）

0 个投票

In my mixed-effect multinomial logistic regression model, the response Y is nominal with more than two levels. The predictors include a random effect (i.e., participant ID), binary predictors, and continuous predictors. The fitglme (<http://www.mathworks.com/help/stats/fitglme.html#namevaluepairs)>, fit generalized linear mixed-effect regression model does not support the multinomial outcome. And I can't find a way to add random effect when using mnrfit (<http://www.mathworks.com/help/stats/mnrfit.html#description)>. Due to the confidentiality of data, I can't post sample data set. Does anybody know if there's a function in MATLAB that can do mixed-effect multinomial logistic regression? Thanks very much!