Seeking Help with Two Small Questions

[manofcrowd]

Dear Professor Wu,

I am a student from China studying your code to determine the etiology scores for pathogens detected in pneumonia patients. Your code has been extremely helpful! I have also read your other literature and code tutorials, but I have encountered two small problems during the replication process:

Question 1: I am not clear about the value of k_subclass. Is k a hypothetical value, or is it obtained based on actual data measurements? Could you provide an example of what the subclass of a pathogen would be in reality?

Question 2: I have studied your tutorial on Model Options for NPLCM with Regression. If I set k to 1, this indicates Non-nested models. However, if my covariates are two discrete variables (AGE GROUP, SITE), then the model options would be as follows:
model_options_reg_nonest$likelihood$Eti_formula = ~ -1 + as.factor(AGE GROUP) + as.factor(SITE)
model_options_reg_nonest$likelihood$FPR_formula = list(MBS1 = ~ -1 + as.factor(AGE GROUP) + as.factor(SITE))
Should the Eti_prior be set to c(2,2) or matrix(1, nrow = N.SITE, ncol = length(cause_list)), or matrix(1, nrow = N.SITE * N.AGE GROUP, ncol = length(cause_list))?

I am a beginner, so my questions might seem foolish—apologies for that. Regardless, thank you for sharing your R package and tutorials. If you believe that answering these questions requires a fee, please let me know. Thank you!

Best regards,