gWQS (version 3.0.5)

gwqs_barplot: Plots and tables functions

Description

Functions that allow to generate plots and tables helping in visualizing and summarise Weighted Quantile Sum (WQS) regression results.

Usage

gwqs_barplot(object, tau, ...)
gwqs_scatterplot(object, ...)
gwqs_fitted_vs_resid(object, sumtype = c("norm", "perc"), ...)
gwqs_levels_scatterplot(object, ...)
gwqs_ROC(object, newdata, sumtype = c("norm", "perc"), ...)
gwqs_boxplot(object, tau, ...)
gwqs_summary_tab(object, sumtype = c("norm", "perc"), ...)
gwqs_weights_tab(object, ...)
selectdatavars(data, na.action, formula, mix_name, ...)
gwqs_rank(data, mix_name, q)

Value

All the plot functions print the output in the Plots pane while the table functions print the output in the Viewer pane.

Qm

The matrix containing the quantiled variables of the elements included in the mixture.

qi

A list of vectors containing the cut points used to determine the quantiled variables.

Arguments

object

An object of class "gwqs" as returned by gwqs.

tau

A number identifying the cutoff for the significant weights. Is tau is missing then reciprocal of the number of elements in the mixture is considered. To avoid printing the threshold line set tau = NULL.

...

Further arguments to be passed.

sumtype

Type of summary statistic to be used: "norm" takes the mean of the estimated parameters on the validation sets and the 95 as the parameters estimates and the 2.5, 97.5 percentiles as CI. This option is only available for objects of class gwqsrh.

newdata

A data frame in which to look for variables with which to predict and generate the ROC curve.

data

Dataset from which you want to select the variables you are interested in.

na.action

Allows to choose what action has to be taken to deal with NAs.

formula

Formula used in the model to specify the dependent and independent variables.

mix_name

Vector containing element names included in the mixture.

q

An integer to specify how mixture variables will be ranked, e.g. in quartiles (q = 4), deciles (q = 10), or percentiles (q = 100).

Author

Stefano Renzetti, Paul Curtin, Allan C Just, Ghalib Bello, Chris Gennings

Details

The gwqs_barplot, gwqs_scatterplot, gwqs_fitted_vs_resid, gwqs_levels_scatterplot, gwqs_ROC and gwqs_boxplot functions produce five figures through the ggplot function.

The gwqs_summary_tab and gwqs_weights_tab functions produce two tables in the viewr pane through the use of the kable and kable_styling functions.

The gwqs_barplot, gwqs_scatterplot plots are available for all family types while gwqs_fitted_vs_resid is not available when family = binomial or "multinomial". gwqs_levels_scatterplot plot is only available when family = "multinomial" and gwqs_ROC when family = binomial. The gwqs_boxplot can be used when the parameter rh within the gwqs function is set greater than 1.

The gwqs_rank function allows to split the variables selected through the vector mix_name in quantiles (depending by the value assigned to q).

Examples

toxic_chems = names(wqs_data)[1:34]
results = gwqs(yLBX ~ wqs, mix_name = toxic_chems, data = wqs_data, q = 4, validation = 0.6,
               b = 2, b1_pos = TRUE, b_constr = FALSE, family = gaussian)

# barplot
gwqs_barplot(results)

# scatterplot
gwqs_scatterplot(results)

# fitted values vs rediduals scatterplot
gwqs_fitted_vs_resid(results)