makeTablePlot: Tabularize the essential result of the predictMe package.

In predictMe: Visualize Individual Prediction Performance

Tabularize the essential result of the predictMe package.

Description

Provides the essential result of the predictMe package, three tables, and, optionally, two plots.

Usage

makeTablePlot(xc = NULL, measColumn = NULL, plot = FALSE, plotCellRes = TRUE)

Arguments

xc	A data.frame with exactly two columns, one of the columns must be the categorized measured outcome, the other column must be the categorized predicted outcome.
measColumn	A single integer number that denotes which of the two columns of function argument 'x' contains the measured outcome.
plot	Logical value, defaults to FALSE. If set TRUE, two complementary plots will be part of the list that this function returns.
plotCellRes	Logical value, defaults to TRUE (is ignored if function argument 'plot' is set FALSE). If set FALSE, the heatmap is returned without frequency results in the cellls.

Details

The c in 'xc' stands for categorized, meaning that the outcome values are expected to have been categorized, so that both columns contain the exact same categories, and are of the class factor.

Columns 1 and 2 of the output 'xTrans' from function binBinary and from function binContinuous provide the expected input of this makeTablePlot function (see Examples).

The returned list will contain 7 items, if function argument 'plot' is set TRUE, if FALSE, it will return the first 5 items (see Values).

Value

a list with five or seven items (see Details):

1. totalCountTable A table with the total counts.

2. rowSumTable A table with proportions that sum up to 1, per row (summing across columns).

3. colSumTable A table with proportions that sum up to 1, per column (summing across rows).

4. rowSumTable_melt The rowSumTable, reformated by the function melt of the reshape2 package.

5. colSumTable_melt The colSumTable, reformated by the function melt of the reshape2 package.

6. rowSumTable_plot The rowSumTable_melt data, plotted by the function ggplot of the ggplot2 package.

7. colSumTable_plot The colSumTable_melt data, plotted by the function ggplot of the ggplot2 package.

Author(s)

Marcel Miché

References

\insertRef

ggplot2WickhampredictMe

\insertRef

reshape2WickhampredictMe

Examples

# Simulate data set with continuous outcome (use all default values)
dfContinuous <- quickSim()
# Use multiple linear regression as algorithm to predict the outcome.
lmRes <- lm(y~x1+x2,data=dfContinuous)
# Extract measured outcome and the predicted outcome (fitted values)
# from the regression output, put both in a data.frame.
lmDf <- data.frame(measOutcome=dfContinuous$y,
                   fitted=lmRes$fitted.values)
# Apply function binBinary
x100c <- binContinuous(x=lmDf, measColumn = 1, binWidth = 20)
# Apply function makeDiffPlot, using columns 1 and 2 from x100c[["xTrans"]]
# The first of columns 1 and 2 contains the measured outcome values.
tp <- makeTablePlot(x100c[["xTrans"]][,1:2], measColumn = 1, plot = TRUE)
# tp is a list with 7 items, items 6 and 7 are the plots that represent
# the numeric information of items 2 and 3 (and 4 and 5, which merely have
# a different format).
# Display item 6 (plot no.1). Perfect performance if the diagonal cells all
# contain the value 1.
tp$rowSumTable_plot

predictMe documentation built on May 24, 2022, 5:05 p.m.