Closed gaow closed 3 years ago
Perhaps, we can add two parameters, z_ld_weight and n_ref, and user must specify one of them. If n_ref is given, the weight is 1/n_ref and this is what we used in simulation. Or we add this recommended weight in the documentation.
I prefer one parameter.
We can recommend they use 1/nref
On Tue, Mar 10, 2020, 11:13 Yuxin Zou notifications@github.com wrote:
Perhaps, we can add two parameters, z_ld_weight and n_ref, and user must specify one of them. If n_ref is given, the weight is 1/n_ref and this is what we used in simulation. Or we add this recommended weight in the documentation.
— You are receiving this because you were mentioned. Reply to this email directly, view it on GitHub https://github.com/stephenslab/susieR/issues/98?email_source=notifications&email_token=AANXRRJV4XGSXAHK3XCTL2TRGZRKRA5CNFSM4LFBUCLKYY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEOMCNNA#issuecomment-597173940, or unsubscribe https://github.com/notifications/unsubscribe-auth/AANXRRLWNEX66EALVABDRSDRGZRKRANCNFSM4LFBUCLA .
I added z_ld_weight
in susie_rss with this commit 33986e2
Separated from #91
@stephens999 I think we have talked about it at some point -- the approach was a bit heuristic and we were not sure what's the right
w
to use. So in a way we never finalized the issue. Are we now at the point where we can comfortably recommend what to do? If not, it may still help to add this parameter to the interface then we can assess it better. I only worry about a formal justification for it that warrants it to be considered a feature.