Graham & Peter meeting
Going over review and figuring out how to address issues. Plan: Will write our line-by-line responses under the reviewer comments in the Google Doc. Peter and I will implement remaining changes on the .tex file, then we’ll compose the reply letter from the notes under the comments and include the latex-diff.
Brief topic notes from today’s meeting:
set up different take on layout: lambda, AIC, power
parameter estimates on over-parameterized
opten implicit nulls
eb bootstrapping is power analysis
rules of thumb only so good, simulate easy.
delta = aic + k
intro on aic use for finding “good enough models” not inferring process
we do stepwise B&K approach (as our recommended approach)
FishBASE interface
Tomomi had an interesting query re FishBASE today - she’d like to be able to sweep through the database and extract certain information such as genus, maximum published weight, etc. While there doesn’t seem to be a documented API per se. theindividual reference pages include a link to an XML summary sheet, where at least some of this information is intelligently tagged. So we can easily set up some R functions with the help of XML and xpath to start grabbing the relevant information.
MPI Computing
Looks like a hardware failure may have been responsible for my crashed MPI job on farm cluster. Waiting to hear more from admin. Likewise waiting on NERSC help on the batch R process on compute nodes. step by step….