plot SPF BinCont: Plots the surrogate predictive function (SPF) in the binary-continuous setting.

Description

Plots the surrogate predictive function (SPF) based on sensitivity-analyis, i.e., \(P(\Delta T | \Delta S \in I[ab])\), in the setting where \(S\) is continuous and \(T\) is a binary endpoint.

Usage

# S3 method for SPF.BinCont
plot(x, Type="Frequency", Col="grey", Main, Xlab=TRUE, ...)

Arguments

x: A fitted object of class SPF.BinCont. See ICA.BinCont.
Type: The type of plot that is requested. The argument Type="Frequency" requests histograms for \(P(\Delta T | \Delta S \in I[ab])\). The argument Type="Percentage" requests relative frequenties for \(P(\Delta T | \Delta S \in I[ab])\). The argument Type="Most.Likely.DeltaT" requests a histogram of the most likely \(\Delta T\) values. For example, when in one run of the sensitivity analysis, \(P(\Delta T =-1| \Delta S \in I[ab])=.6\), \(P(\Delta T =0| \Delta S \in I[ab])=.3\), and \(P(\Delta T =-1| \Delta S \in I[ab])=.1\), the most likely outcome in this run would be \(P(\Delta T =-1\). The argument Type="Most.Likely.DeltaT" generates a plot with percentages for the most likely \(P(\Delta T)\) value across all obtained values in the sensitivity analysis.
Col: The color of the bins or lines when histograms or density plots are requested. Default "grey".
Main: The title of the plot.
Xlab: Logical. Should labels on the X-axis be shown? Default Xlab=TRUE.
...: Arguments to be passed to the plot, histogram, ... functions.

Author

Wim Van der Elst & Ariel Alonso

References

Alonso, A., Van der Elst, W., Molenberghs, G., & Verbeke, G. (2017). Assessing the predictive value of a continuous surogate for a binary true endpoint based on causal inference.

Examples

Run this code

if (FALSE)  # time consuming code part
data(Schizo_BinCont)
# Use ICA.BinCont to examine surrogacy
Result_BinCont <- ICA.BinCont(M = 1000, Dataset = Schizo_BinCont,
Surr = PANSS, True = CGI_Bin, Treat=Treat, Diff.Sigma=TRUE)

# Obtain SPF
Fit <- SPF.BinCont(x=Result_BinCont, a = -30, b = -3)

# examine results
summary(Fit1)
plot(Fit1)

plot(Fit1, Type="Most.Likely.DeltaT")