Same model set, multiple species

questions concerning analysis/theory using program PRESENCE

Same model set, multiple species

Postby northernbio » Thu Nov 18, 2010 4:21 pm

I would like to create a common set of models that I will use for analysis of detection probabilities for about 14 different species (I am using the Presence 3.1), and would like to partially automate the model specification steps, but am having trouble understanding the proper way to do this. I believe I should be using "Analysis from file". Would I first create my set of models for one of the species, then copy and paste the .pa3 file into project directories for each of the other species? Are there additional steps that are needed?
northernbio
 
Posts: 20
Joined: Thu Nov 18, 2010 11:08 am

Re: Same model set, multiple species

Postby darryl » Thu Nov 18, 2010 4:42 pm

Not sure if Jim Hines has fully updated this feature or not with the newest version, but the basic intent is that you fit all the models of interest to the data for 1 species. You then create a new project for the next species and select the 'analysis from file' option. You will then be prompted to select the old project that contains the models you want to fit to the new data, and then asked to select specific models. Click 'run' and viola! All done. At least in theory....

Now you have to make sure that you have exactly the same covariate names in both files, and the same number of surveys. You also need to go through the output of each model to check for any warnings; don't just look at the resulting model summary table!

Hope that helps
darryl
 
Posts: 498
Joined: Thu Jun 12, 2003 3:04 pm
Location: Dunedin, New Zealand

Re: Same model set, multiple species

Postby matobler » Thu Nov 18, 2010 9:21 pm

I was facing the same problem a while back and after playing around with the 'analysis from file' function in Presence (2.0 at that time) and constantly running into problems (it does/did not work for the RN and other models) I gave up and switched to RMark. Once you have your code setup in R, it is just a matter of changing the data file name to run the same set of models for different species. So I strongly encourage you to take a look at RMark. If you are already familiar with R and linear models, running the models in RMark will be pretty straight forward, if not your time is probably still better spent figuring this out in RMark than in Presence. I would be happy to send you some code examples.

Mathias
matobler
 
Posts: 19
Joined: Fri Nov 03, 2006 9:44 pm
Location: Peru

Re: Same model set, multiple species

Postby jhines » Fri Nov 19, 2010 12:05 am

The procedure is to create a project folder for each species, build/run a set of models for one species, then use the 'Run analysis from file' menu item to run the models for the other species. After creating the first species project file and running your models, exit PRESENCE and restart and build the project folder for the 2nd species. Then use 'Run analysis from file' and click 'select all'. When it finishes, exit PRESENCE, restart and do the 3rd species. etc.

Jim
jhines
 
Posts: 632
Joined: Fri May 16, 2003 9:24 am
Location: Laurel, MD, USA

Re: Same model set, multiple species

Postby northernbio » Fri Nov 19, 2010 11:50 am

Awesome, thanks for the replies. One more related automation issue. My site names and covariates are of course the same for each datafile, and I see there is a function called "Insert from CSV" table. Can I use this function for inserting covariates, and if so, what are the field requirements (e.g., do I need field headers and a column of site names, and can I add more than 1 covariate from a single .csv file)?
northernbio
 
Posts: 20
Joined: Thu Nov 18, 2010 11:08 am


Return to analysis help

Who is online

Users browsing this forum: No registered users and 4 guests