Sunday
Sent in my review for PLoS ONE. No way to attach pdf files, had to create and send a secure link instead to provide comments on the pdf.
AIC discussion on r-sig-phylo. For a derivation I refer to Cavanaugh, (1997), also see Daniel Schmidt’s slides. Aikaike’s original paper does the derivation less deliberately than Cavanaugh but does discuss this Frequentist intepretation as the maximum expected log likelihood (Akaike, 1974).
Monday
- Discussing peerJ, NCEAS publishing working group
- knitcitations submitted to CRAN
- Check fallacy appendix
Runs for Value of Information:
Stationary policy function and value function plots for the seven noise combinations, with Beverton-Holt dynamics.
I also re-ran this with strongly nonlinear recruitment function for comparison (Ricker model with allee threshold below which stocks crash): Allee dynamics
Value of information meeting (Jim + Mike)
- check surprising result in growth noise using negative shocks only. Done: confirms expected behavior.
- run with Sethi exact settings (overnight, results)
- Think about Mike’s point in differences between parameter uncertainty and stochasticity. See cost_of_bias.md results as a starting point.
Tuesday
- check fallacy appendix
- edits to knitcitations from feedback
- Use github url for all package pages on CRAN
- Submitted fallacy paper!
Wednesday
Value of information
- debugging Sethi exact settings. Try running logistic + uniform shocks on Reed-only noise.
Evolution stuff
- treebase finalize edits?
- ievobio submission?
- wrightscape edits
Misc CSS notes
CSS can’t replace LaTeX yet, but it’s getting most of the basics of good typography. We might at least have LaTeX quality justified text soon, see:
References
- Cavanaugh JE (1997). “Unifying The Derivations For The Akaike And Corrected Akaike Information Criteria.” Statistics & Probability Letters, 33. ISSN 01677152, https://dx.doi.org/10.1016/S0167-7152(96)00128-9.
Akaike H (1974). “A New Look at The Statistical Model Identification.” Ieee Transactions on Automatic Control, 19. ISSN 0018-9286,