pseudo R2 is a sort of measure of model fit comparing the full model to the null model, so you may want to check other fit stats, for example, computing the squared correlation between observed and predicted values?
see if this article is helpful?
http://www.ats.ucla.edu/stat/sta ... pter3/statalog3.htm
扫码加好友，拉您进群
全部版块
我的主页
收藏
