v1.1.2

matrix_model had some issues with sparse matrix input (matrices with many zeroes) that have now been resolved.

colorednoise v1.1.1

A patch release to deal with the @docType package error in roxygen2.

colorednoise v1.1.0

• New function stdev_transform, which "translates" standard deviation from the normal distribution to lognormal or logit
• The data handling backend has been changed from dplyr to data.table for faster simulations.

colorednoise v1.0.5

• Updated to reflect changes in tibble v2.0.0
• Updated to reflect changes in tidyr v1.0.0

v1.0.3

matrix_model now has more options for dealing with survival values erroneously set to >1.

v1.0.2

• matrix_model now automatically detects structural zeroes in matrix input.
• The Makevars files have been tweaked so that the package will work with more operating systems and compilers.
• colorednoise no longer depends on the deprecated package purrrlyr.

v1.0.0

colorednoise can now run matrix population models with colored noise. Not only that, but it runs matrix population models, with the option of demographic stochasticity, faster than any other R package implementation I've tested. This version is not backwards-compatible because I've given the raw_noise and timeseries functions much more intuitive and sensible names: colored_noise and unstructured_pop.

v0.0.2

The simulations in version 0.0.2 run three times faster thanks to recoding in Rcpp. The new version supports separate autocorrelation values for survival and fertility in population time series, and the variance in survival probabilities should now match expectations.