Varying AIC

[Daniel28052]

Hi,

I am using tidyLPA to identify the optimum number of latent profiles in my data set. I am aware that several informtion criteria can be applied looking for the best model (such as the Akaike Information Criteria; AIC). However, I am interesting to know why does the AIC vary when calculated multiple times with the same number of indicators and same number of estimated profiles.

I was wondering if someone can help me with this.

Thank you!!
Daniel.