version 1.2.0
* Function that can return desired asymmetric cost models
* Alphabetical order of coefficients for the inference summary function
* Big speedup to confusion_results function via a custom C++ table routine and an option to perform argument appropriateness checks
* Small bug fix when inference is not desired
* Small bug fix when proportions_scaled_by_column = TRUE in the general_confusion_results function

version 1.1.0
* Allows passing in of a custom matrix multiplication function and a custom square root of the diagonal of a matrix inverse thereby giving the opportunity to use CUDA / GPUs (see README on github for an example usage)
* Includes an algorithm for inference of only one coefficient using conjugate gradient descent which is faster than inference-for-all coefficients for some settings

version 1.0.1
* Removed OMP dependency to please CRAN

version 1.0.0
* Initial Release