Agree on model name?

[athowes]

Goal

We should choose a shorter name for the "latent truncated censoring adjusted delay" model. This name would also be used as a class modelname_epidist.

Context

• Originally called latent_truncated_censoring_adjusted_delay
• I have abbreviated to ltcad in S3 structure refactor #70
• In thinking about this, we should consider whether the package will contain any models which are not "truncatation adjusted" or "censoring adjusted" i.e. how much does this need saying?

Required features

• A shorter name
• It could be descriptive, it could not
• It should be compatible with our future model development interests (i.e. distinguishable from future models we might want to include)

[seabbs]

I think I vote that we assume (i.e unless we communicate otherwise) that all methods will account for right truncation and censoring at least and so this doesn't need to be explicitly stated.

I think that due to this calling this latent would make sense.

We could also call it after the paper it comes from (here ward) but that might get messy with multiple methods per paper (e.g. park)

[athowes]

Agree that focus should be on methods should account for right truncation and censoring.

Wonder if any methods that don't may be eventually included as teaching example. If so, could potentially use some naming to distinguish that they don't account for (as in, accounting is default).

Calling "latent" doesn't really seem like a name to me. Latent variable model? Hierarchical model? Latent state model?

I think calling it after the paper isn't a very extensible strategy. Especially if we are expecting to be developing methods (which won't then have names).

[athowes]

@kgostic on call:

• Latent one acts act the individual level
• Down the road perhaps a delay one (group level)
• Think about which models we will have in future

@seabbs:

• Won't know which models we will have in the future
• individual_latent_variable seems reasonable

In theory could extract individual times for primary and secondary events.

People seem happy to be moving forward with this.

[seabbs]

Update from call that most people like latent_individual (or vice versa).