Very large likelihoods

[Alexis Cerezo]

Dear all,

I'm running an analysis in PRESENCE with bird presence-absence data, in three habitats (tropical forest and two agroforestry habitats, pepper/cinnamon and cardamom) and 4 years of data.  I ran the following models:

NUM Model
1 psi(.),p(.)
2 psi(year),p(.)
3 psi(habitat),p(.)
4 psi(habitat+year),p(.)
5 psi(year),p(habitat)
6 psi(habitat),p(habitat)
7 psi(habitat+year),p(habitat)
8 psi(habitatxyear),p(.)
9 psi(habitatxyear),p(habitat),

with the following B's:

for psi:

B1-ordinate (represents tropical forest)
B2-pepper/cinnamon effect
B3-cardamom effect
B4-year
B5-year x pepper/cinnamon effect
B6-year  cardamom effect

for p:

B1-ordinate (represents forest)
B2-pepper cinnamon effect
B3-cardamom effect

The results file at first looked ok, including the Bi estimates for each model, but when I looked at the model likelihoods, for some models they were HUGE!! my question is thus, are these values correct, or is there a problem?

Here is an example from the PRESENCE output:

Model                                    AIC       Model Likelihood
psi(.),p(.)                            438.24       1.00
psi(year),p(.)                            439.90       0.44
psi(habitat),p(.)                    416.35       56669.99
psi(habitat+year),p(.)            417.89       26238.94
psi(year),p(habitat)            420.83       6033.00
psi(habitat),p(habitat)            411.40       673336.17
psi(habitatxyear),p(.)            414.67      131268.48
psi(habitat+year),p(habitat)  412.67       356824.73
psi(habitatxyear),p(habitat)  409.76       1528809.67

(see likelihoods for models 3-9)

Any help will be greatly appreciated, cheers.

Alexis

[Marc Kery]

Dear Alexis,

if I knew an answer I'd let you know .... but anyway, it'd be safer for you to ask on the specific email list for PRESENCE. Please see here:http://www.phidot.org/forum/viewforum.php?f=14

Best regards  --- Marc

---

From: unma...@googlegroups.com <unma...@googlegroups.com> on behalf of Alexis Cerezo <alexis...@gmail.com>
Sent: Tuesday, June 18, 2024 15:49
To: unmarked <unma...@googlegroups.com>
Subject: [unmarked] Very large likelihoods

 

--
*** Three hierarchical modeling email lists ***
(1) unmarked (this list): for questions specific to the R package unmarked
(2) SCR: for design and Bayesian or non-bayesian analysis of spatial capture-recapture
(3) HMecology: for everything else, especially material covered in the books by Royle & Dorazio (2008), Kéry & Schaub (2012), Kéry & Royle (2016, 2021) and Schaub & Kéry (2022)
---
You received this message because you are subscribed to the Google Groups "unmarked" group.
To unsubscribe from this group and stop receiving emails from it, send an email to unmarked+u...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/unmarked/e5e8fac9-59d4-4e2c-9761-ae63cd18ee04n%40googlegroups.com.

[Alexis Cerezo]

Hello Marc,

Yes, unfortunately I'm having problems with the phidot page, but I´ll send my query there when I figure out what's going on.

Many thanks and best regards,

Alexis

To view this discussion on the web visit https://groups.google.com/d/msgid/unmarked/ZR0P278MB0869344613236A634BA54F1EEBCE2%40ZR0P278MB0869.CHEP278.PROD.OUTLOOK.COM.