search

rbchan / unmarked

R package for hierarchical models in ecological research
https://rbchan.github.io/unmarked/
37 stars 25 forks source link

Can't make fitList from occuFP models #147

Closed kenkellner closed 4 years ago

kenkellner commented 4 years ago

Probably need to convert fitList() into a method instead of a regular function.

rbchan commented 4 years ago

Good idea

rbchan commented 4 years ago

Andy, did you get around to submitting Ken's changes to CRAN? If you're too busy, do you want me to do it?

On Wed, Oct 2, 2019 at 7:06 PM Ken Kellner notifications@github.com wrote:

Probably need to convert fitList() into a method instead of a regular function.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/rbchan/unmarked/issues/147?email_source=notifications&email_token=AABWRCQ4M6532EMW2MY4LADQMUSNVA5CNFSM4I44NR62YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HPIMTTQ, or mute the thread https://github.com/notifications/unsubscribe-auth/AABWRCXSDKVZXRFGNDQSQTTQMUSNVANCNFSM4I44NR6Q .

rbchan commented 4 years ago

Hey man , Sorry about that ... if you have time that would be great ... on my way home rider now and traveling again the next 2 weeks I’m afraid ... just no time lately. Sorry man!

Great seeing you this week , wish I could have spent the whole week , what a great meeting! Regards Andy

Sent from my iPhone

On Oct 3, 2019, at 12:35 PM, Richard Chandler richard.chandlers@gmail.com wrote:

Andy, did you get around to submitting Ken's changes to CRAN? If you're too busy, do you want me to do it?

On Wed, Oct 2, 2019 at 7:06 PM Ken Kellner notifications@github.com wrote:

Probably need to convert fitList() into a method instead of a regular function.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/rbchan/unmarked/issues/147?email_source=notifications&email_token=AABWRCQ4M6532EMW2MY4LADQMUSNVA5CNFSM4I44NR62YY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4HPIMTTQ, or mute the thread https://github.com/notifications/unsubscribe-auth/AABWRCXSDKVZXRFGNDQSQTTQMUSNVANCNFSM4I44NR6Q .

kenkellner commented 4 years ago

I am working on a patch that's almost done for automated model selection with all covariate subsets (like MuMIn::dredge()) inspired by Dana Morin's talk. If you want to wait for that and the adjustments to fitList I will probably get them done in the next week.

rbchan commented 4 years ago

Ken, I'm a bit concerned that "all subsets" approaches can increase the chance of finding spurious results. Is there a strong need to add that functionality if it already exists in other packages?

Sorry I didn't get a chance to meet you at the meeting yesterday. Next time!

On Thu, Oct 3, 2019, 12:28 PM Ken Kellner notifications@github.com wrote:

I am working on a patch that's almost done for automated model selection with all covariate subsets (like MuMIn::dredge()) inspired by Dana Morin's talk. If you want to wait for that and the adjustments to fitList I will probably get them done in the next week.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/rbchan/unmarked/issues/147?email_source=notifications&email_token=AABWRCT2VCBMXYLEIAZVSNTQMZBUZA5CNFSM4I44NR62YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEAJJW7A#issuecomment-538090364, or mute the thread https://github.com/notifications/unsubscribe-auth/AABWRCTDREJ6MF4ESRDBJ3TQMZBUZANCNFSM4I44NR6Q .

kenkellner commented 4 years ago

Well, I don't think there's any existing approach that works with unmarked models (given the unique multi-formula structure). But I understand the concern about questionable results.

The dredge stuff is really just a wrapper on a function for fitting a bunch of specific models together / in parallel. That at least seems worth including to me since many people seem to be fitting large numbers of models based on different hypotheses etc. and there currently isn't a quick way to do that.

rbchan commented 4 years ago

Ok, sounds good. Will you send the tar ball to Andy and me when it's ready? Andy will upload it to cran.

On Thu, Oct 3, 2019, 1:55 PM Ken Kellner notifications@github.com wrote:

Well, I don't think there's any existing approach that works with unmarked models (given the unique multi-formula structure). But I understand the concern about questionable results.

The dredge stuff is really just a wrapper on a function for fitting a bunch of specific models together / in parallel. That at least seems worth including to me since many people seem to be fitting large numbers of models based on different hypotheses etc. and there currently isn't a quick way to do that.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub https://github.com/rbchan/unmarked/issues/147?email_source=notifications&email_token=AABWRCURY3NYE23SF4ZME7TQMZEZTA5CNFSM4I44NR62YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEAJMVTQ#issuecomment-538102478, or mute the thread https://github.com/notifications/unsubscribe-auth/AABWRCXRZH6GCAG23R4HKOLQMZEZTANCNFSM4I44NR6Q .

jaroyle commented 4 years ago

hi fellas, I've been sicker than a dog since returning from TWS and hope to have my head back above water later this week. Sorry to be out of touch. If someone can send me the tarball I will upload it so as to not have to change maintainer. Sorry to drop the ball on this but I have a trip next week that I have to focus all burners on right now... regards andy

On Thu, Oct 3, 2019 at 6:36 PM Richard Chandler notifications@github.com wrote:

Ok, sounds good. Will you send the tar ball to Andy and me when it's ready? Andy will upload it to cran.

On Thu, Oct 3, 2019, 1:55 PM Ken Kellner notifications@github.com wrote:

Well, I don't think there's any existing approach that works with unmarked models (given the unique multi-formula structure). But I understand the concern about questionable results.

The dredge stuff is really just a wrapper on a function for fitting a bunch of specific models together / in parallel. That at least seems worth including to me since many people seem to be fitting large numbers of models based on different hypotheses etc. and there currently isn't a quick way to do that.

— You are receiving this because you commented. Reply to this email directly, view it on GitHub < https://github.com/rbchan/unmarked/issues/147?email_source=notifications&email_token=AABWRCURY3NYE23SF4ZME7TQMZEZTA5CNFSM4I44NR62YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEAJMVTQ#issuecomment-538102478 , or mute the thread < https://github.com/notifications/unsubscribe-auth/AABWRCXRZH6GCAG23R4HKOLQMZEZTANCNFSM4I44NR6Q

.

— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/rbchan/unmarked/issues/147?email_source=notifications&email_token=AAHGZC4ZAFJPCEB2ASXT62TQMZXXNA5CNFSM4I44NR62YY3PNVWWK3TUL52HS4DFVREXG43VMVBW63LNMVXHJKTDN5WW2ZLOORPWSZGOEAJZTXA#issuecomment-538155484, or mute the thread https://github.com/notifications/unsubscribe-auth/AAHGZCZM5S3EY24MB3BKQ6LQMZXXNANCNFSM4I44NR6Q .