Assembling material/outline for my SICB presentation in January, will be giving a practice talk in the Wainwright lab on Wednesday. 15 minute presentation with 5 minutes for questions.
Designing presentation in “storyboard” format in Inkscape. Will figure out if I convert to slides or not.
Simulations under the Anoles dataset to show posterior distributions of lambda. Exploring the connection between the likelihood of a parameter under each model to that of the distributions of the likelihood ratio. Clearly more valuable to look at likelihood of the whole data than that of a (not necessarily good) summary statistic (model parameter that isn’t a sufficient statistic).
To that effect, I’ve added the ability to estimate likelihood of a parameter under the alternative model, rather than just bootstrapping parameters directly.
Have also begun developing socialR package to handle the social reporting functions more smoothly.
See github and flickr logs for today for more details.