Research

  • Working on bootstrapping theta parameters
  • bootstrapping multiple parameters doesn’t seem effective in the context of the boot package, despite claiming it can use a vector of stats. Certainly part of challenge is that different models have different numbers of parameters, so this isn’t a fixed size vector. Approaching manually for the moment, inhibits use of bootstrap confidence intervals function from the boot package.
  • Some irregular behavior in bootstrapping parameters – have to load package and then reload likelihood_bootstrap library source file or else bootstraps are NAs, needs further investigation.
  • formatting bootstrapping meaningfully – following the ouch style for the moment
  • Sent Peter Wainwright a rough draft of Bio Let. manuscript.

Goals / ToDo

  • Add bootstraps of parameters to manuscript,
  • Function to bootstrap all parameters given list of all models. Plotting function also.
  • Write up Evolution note on AIC.
  • Conference registrations
  • NESCent whitepaper

Primates

Reading & Notes

  • Embedding google spreadsheets sounds much more promising than formatting tables in OWW/html: Example from Anthony / Koch Lab
  • OWW Stats
  • FF discussion on Chelsea’s Open Science article
  • Embedding/Transclusion in OWW/mediawiki and examples

  • Exploring gist for code fragments. Embeddable gist: