A new function has been added. The function is currently in its alpha phase:
With this variant of the xy_plot function one can visualise interactions between two parametric predictors in the compact style of xy_plot. The function can be used and should work in most if not all cases, but keep in mind that it is still in an early phase of development.
The following function is no longer in its alpha phase:
Additionally, a qualitify of life improving modification to the following function was made:
Facet labels now can be set manually using the newly introduced names argument. See the reference for details.
A new function has been added. The function is currently in its alpha phase:
This function creates a point plot with confidence interval ranges using ggplot2. It combines the model estimates for the same predictor in two (sets of) QGAMs, i.e. it displays estimates of X and Y coordinates. It basically is a combination of two parametric plots, with estimates for X coordinates plotted on the X axis, and estimates for Y coordinates plotted on the Y axis. The function can be used and should work in most if not all cases, but keep in mind that it is still in an early phase of development.
One minor change has been added to:
The individual summaries are now separated more clearly.
Two functions have been added:
The extract_xyt function is a wrapper of the three already included extraction functions. With the new S3 method for the object class mqgam, mqgam objects now can be summarized.
Four data objects have been added:
The data objects are used to run examples. They are ‘lazy loaded’ when loading the mtqgam package, and can be used by users to test functions.
Additionally, qualitify of life improving modifications to the following functions were made:
Both plot functions now correctly work with predictors of more than two levels. Previously, a coding error made specifying colours necessary. The extract functions now take mousetrap objects as data source. Previously, data extracted from mousetrap objects had to be specified. While this is still possible, I advise to use the newly created argument usage. Please see the examples in the pertinent references for further details.
A new function has been added:
The function creates plots for factor:factor and numeric:factor interactions in QGAMs. For factor:factor interactions, two facets of dot plots with confidence interval ranges are plotted. For numeric:factor interactions, a smooth plot with confidence intervals is plotted. The smooth plot is either given in two facets (default) or as one plot (see the type argument in the documentation/reference for further information).
Three functions have been added:
Additionally, qualitify of life improving modifications to the following functions were made:
That is, both plotting functions now take a much more simple pred argument. Before, this had to be a named list of factor levels; now, this is just the predictor variable itself. Other modifications include the addition of more meaningful error messages and streamlining the argument structure overall.
The extract_ functions now display a progress bar when run. This can be toggled off specifying the newly introduced verbose argument as false.
Code created with prior versions of the package continues to function. Please have a look at the references and the documentation of the pertinent functions for more detailed information.
One major change to the following function is part of this version:
The function now recognises the cond argument for choosing between interacting terms to display.
This is the initial version of the mtqgam package. The following functions are included: