NEWS.md

In levisc8/spind: Spatial Methods and Indices

spind v2.2.1

• Removed unneeded functions from Namespace

• Makes WRM a lot noisier when the model breaks.

spind v2.2.0

• deprecated customize_plot and plot arguments in favor of new plot methods for GEE and WRM.

  • GEE and WRM now return the ggplot object in their respective classes. Users can modify them from there rather than use the clunky interface to modify them within the model function call.

• rvi.plot and th.indep now return ggplot as list items along with the values that they returned before. customize_plot arguments have been deprecated in every function it appears in. plot.ROC in th.indep is also deprecated.

• covar.plot's customize_plot argument is also deprecated, but the plot argument is not, as this controls which statistic is plotted (variance or covariance) rather than whether or not a plot is generated at all. This function no longer prints the ggplot object by default, but returns it as part of a list.

• deprecated color.maps arguments to upscale and adjusted.actuals.

• Some documentation clarification and typo fixes.

• Updated a couple error messages for consistency across functions.

• Behind the scenes, moved all plotting functions that use ggplot2 to tidy evaluation equivalents. If you're interested, more on that here.

spind v2.1.3

-Updated citation information.

spind v2.1.2

-Updated plot outputs from GEE and WRM so that y-axes aren't absurdly packed when autocorrelation values are very large -Updated documentation for some modeling functions to reflect that factors as expanatory variables are not supported by spind -Added some continuous integration functionality to the development branch to ensure that bugs are caught faster. No package functionality is affected by this

spind v2.1.1

-fix bug in predict.WRM that prevented calculating smooth components -fix bug in scaleWMRR that prevented acfft from being called internally -added trace argument to mmiGEE -updated vignette to teach users how to customize plots

spind v2.1.0

-Updated step.spind model hierarchy recognition algorithm for improved efficiency and generality. Additionally reformatted code so no error message is produced when the initial model is the best model. -Fixed examples in wavevar and wavecovar which called a non-existant function. -Updated the package vignette for clarity and formatting. Additionally, removed pointless WRM example (padding with mean values) because the particular case being demonstrated was identical to first example. -Added prettier plots from GEE, WRM, covar.plot, rvi.plot, and th.indep using ggplot2 graphics. These functions should produce publication quality graphics, rather than the bare bones, minimalist approach in the previous versions of the package. -Provide optional color palettes for graphical outputs of adjusted.actuals and upscale so they are more aesthetically pleasing. -Updated argument names for a couple functions to ensure consistency throughout the package.

spind v2.0.1

-Updated vignette to newest version. Old version is not well developed. -added citation() to package

spind v2.0.0

introduces functions for implementing GEEs and WRMs in the context of spatial models in R. We are also including multiple model selection options for you to find the best model possible for your data. We have retained all of the original functions to calculate spatially corrected measures of model fit.

GEEs

We've added wrappers for the gee and geese functions found in packages gee and geepack so that they can be implemented for spatial models. We've also added S3 methods for summary and predict to make them easy to interact with.

WRMs

In addition to wrappers for GEEs, we've added wrappers for functions found in the waveslim package that allow you to implement wavelet-revised models on spatial data sets. As with GEEs, we've also added S3 methods for predict and summary.

Multi-model inference and model selection

We've added functions to conduct multi-model inference and stepwise model selection (backwards only for the time being), as well as a slew of helper and utility functions to examine your model in greater detail.

Previous functionality

The spatial indices included in the original release should still work as they did before.

levisc8/spind documentation built on April 3, 2024, 4:52 a.m.