JOSS release

Release for completion of review for Journal of Open Source Software (JOSS). Updates included minor bug fixes, code standardization, major updates to documentation, addition of a vignette, and coverage of unit testing.

Model evaluation functionality release

This is a functionality update release of the klrfome package. The name is an acronym for Kernel Logistic Regression on FOcal Mean Embedding Kernels (pronounced as "clear foam"). This package computes a mean embedding similarity kernel, performs kernel logistic regression, and predicts on raster images with the option for parallel processing. This process is developed with the express interest in predicting archaeological site sensitivity, but the model is likely useful for other applications within and beyond archaeology. This release includes a series of functions for evaluating the performance of model predictions using a variety of common (and uncommon) metrics. These metrics are those derivable from the quantities of a confusion matrix. Additional updates include bug fixes and an updated Dockerfile for the installation of the NLMR package. The package README includes a fully reproducible workflow using simulated data and includes an example of the new evaluation functions. This release is published following the presentation of this package at the SAA 2018 meeting in Washington, D.C. (April 12th, 2018). Please file issues or pull requests if bugs are encountered.

Initial Release

This is the initial release of the klrfome package (previously names DistRegLMERR). The name is an acronym for Kernel Logistic Regression on FOcal Mean Embedding Kernels (pronounced as "clear foam"). This package computes a mean embedding similarity kernel, performs kernel logistic regression, and predicts on raster images with the option for parallel processing. This process is developed with the express interest in predicting archaeological site sensitivity, but the model is likely useful for other applications within and beyond archaeology. This initial release contains the full suite of functions to perform the stated data preparation, model fitting, and prediction. This package README includes a fully reproducible workflow using simulated data. As an initial release (in preparation for the SAA 2018 conference, Washington, D.C.), there will be bugs and the will be changes in the future. Please file issues or pull requests as needed.

initial release to Zenodo

Initial release: R package builds, mostly working functions, so example code. Still needs a ton of work. This release is to get it on Zenodo