Online repository to reproduce the results presented in :
Fraser Franco, M., Santostefano, F., Martin, G. A., Julien, Kelly, Clint D., Montiglio, P.-O. (XXXX) Individual foraging specialization and success change across experience in a virtual predator-prey system.
The subfolders in the code folder are enumerated in order to reflect the worflow.
Here are the steps highlighting the process with a link to the code subfolders :
You will find the specific outputs generated from the R scripts in the outputs folder. Each output file has a number that correponds to the R script used to generate it.
I ran all the models on Cedar, a computer cluster managed by the Digital Research Alliance of Canada. You can open the .sh files in the jobs_slurm folder to see the computer specifications used to run the models. The computer cluster runs on CentOS Linux 7.
Here is a table showing the workflow employed to generate the model outputs using R. This workflow is exclusively for .R files used to compute the Bayesian mixed models with the package "brms" (see the 01_code_models folder).
| Generator | Operating system | Programming language | Code | Results |
|---|---|---|---|---|
| Digital Research Alliance of Canada / Cedar | CentOS Linux 7 | R | code folder | outputs folder |
The model fitting can be reproduced on your personal computer but requires important computing resources so fitting may take some time.
Output files will be promptly uploaded on an OSF repository. Stay tuned...
I used the renv package to contain each individual R package version within the project. If you wish to reproduce our results in your personal computer (excluding the model files that were run on Cedar), please refer to the official renv collaborating vignette to implemement the workflow on your personal computer.