I am working a causal inference problem with longitudinal data, and I find this interesting package and hope to use it in my data analysis. I have a lot of time points and the default method for $g$ and $Q$ would consider all the parent variables into the models, which gives me some problems. I am wondering how to force the $g$ and $Q$ function only depend on the baseline covariates and the most recent $L_nodes$. I check your help function for deterministic.Q/g.function, but cannot understand it well.
Hi,
I am working a causal inference problem with longitudinal data, and I find this interesting package and hope to use it in my data analysis. I have a lot of time points and the default method for $g$ and $Q$ would consider all the parent variables into the models, which gives me some problems. I am wondering how to force the $g$ and $Q$ function only depend on the baseline covariates and the most recent $L_nodes$. I check your help function for deterministic.Q/g.function, but cannot understand it well.
Hope to see your reply.
Best,
Hao