coencliner Change Log
Version 0.2-2
* Fix bug in `NegBin()` and `ZINB()` which had incorrect
specification of the gamma mixture part of the distribution.
Reported by: @nlhuong #24
* Allow for vector `alpha` in `NegBin()` and `ZINB()`
Version 0.2-1
* Fix problems identified by `R CMD check` under R-devel
Version 0.2-0
* Bug Fix: A typo in the bivariate Beta response function led
to the code ignoring `gamma` for the second gradient, which
used the value for first gradient instead.
Reported by: Jari Oksanen (with patch/fix)
* For additional details, please see the github commit log:
http://github.com/gavinsimpson/coenocliner/commits/master
For higher-level news of chanegs to the package, see the NEWS
file.
Version 0.1-0
* Released to CRAN...
Version 0.0-10
* Gaussian: default for `corr` changed to `0` so users don't
need to specify this unless they want correlated responses.
* vignette: added a basic tutorial vignette document to the
package.
* BetaBinomial: realised I should have asked Jari for
permission to copy code to compute the parameters of the
beta distribution from a parameter tau^2. This came from his
BETASIMU C code. I've temporarily removed this code until such
a time as I get permission to include it (Jari is in the field).
I've replaced it with a different parameterisation I found in
Ben Bolker's Ecological Models and Data in R book.
Version 0.0-9
* ZIP, ZINB: new wrappers for random number generation for the
zero-inflated Poisson (ZIP) and Negative binomial (ZINB)
distributions. These functions are simple ZIP and ZINB
distributions where the probability of a zero from the binomial
part of the model depends only on a mean probability of zero.
Version 0.0-8
* Binomial, BetaBinomial: New wrappers for random number
generation for the Binomial and Beta-Binomial distributions.
The Binomial distribution allows simulation of binomial counts
from a probability of occurence and binomial denominator, m,
the number of trials (number of individuals counted).
The Beta-Binomial() is an extra-variance Binomial distribution
and allows simulation of overdispersed binomial count data in a
manner similar to that by which the Negative binomial
distribution can be used to generate overdispersed Poisson
count data.
Version 0.0-7
* Beta: now implemented for one or two gradients.
Version 0.0-6
* Binomial(): this really should have been Bernoulli and now is.
Version 0.0-5
* Old code: This version removed all the `simxDfoo()` functions
and associated helpers (`betaResponse()`, `gaussianResponse()`,
`biGaussianResponse()`, `expandGauss()`, `expandBeta()`) as
these are now no longer needed given that we have `coenocline()`
as a gneral interface to all these things.
Version 0.0-4
* Gaussian() & Beta() modified to take lists of arguments, `px`
and `py`, to simplify the naming of species parameters in these
models.
* coenocline() modified to supply arguments in the format now
required by `Gaussian()` and `Beta()`.
Version 0.0-3
* coenocline() a generic interface to coenocline simulation.
* Beta(), Gaussian(); new response model functions for the
classic Gaussian response model and the generalised Beta
response model.
* New wrappers for random deviate generation from Poisson,
Binomial, and Negative binomial distributions.
* expand() a new general version of the expand.grid()-like
functionality where we repeat sets of parameters for each of
n gradient locations.
Version 0.0-2
* added simulators for occurrence not abundance
Code provided by F. Rodriguez-Sanchez.
Version 0.0-1
* intial alpha version