how to specify the covariance structure across groups
Hi~
I'm using MEFISTO to analyze a longitudinal dataset with approximately 100 subjects, each having multiple time points (ranging from 2 to 10). However, the computation time is excessively long. I would like to simplify the covariance structure to improve efficiency, but I haven't found relevant parameters for this adjustment. Could you provide guidance on how to achieve this or suggest alternative ways to speed up the analysis?
So far, I have tried filtering features, using stochastic inference, and enabling use_float32, but the performance is still an issue. Any insights would be greatly appreciated. Thank you!
You can use mefisto_opts$model_groups = FALSE to speed up computation and not perform optimization of the group covariance structure.