simone (version 1.0-4)

plot.simone: Graphical representation of SIMoNe outputs

Description

Plots various outputs associated to a SIMoNe run.

Usage

# S3 method for simone
plot(x,
     output    = c("BIC", "AIC", "ROC", "PR", "path.edges",
                   "path.penalty", "sequence"),
     ref.graph = NULL,
     ask       = TRUE,  ...)

Arguments

x

output of a simone run (must be an object of class simone)

output

a vector of character string indicating which outputs must be plotted (picken from "BIC", "AIC", "ROC", "PR", "path.edges", "path.penalty" or "sequence"). Default is to plot everything possible.

ref.graph

a network of reference provided through an adjacency matrix that is used to compute the ROC and PR curves. Only required if "ROC" and "PR" belongs to the output argument.

ask

a logical indicating if the graphics device should be interactive. Default is TRUE.

…

Additonal arguments for generic plot (such as main = "my title").

Details

Here are some details about the plots possibly produced:

• If "BIC" belongs to the output argument, a plot representing the Bayesian Information Criterion as a function of each network inferred by simone is displayed.

• If "AIC" belongs to the output argument, a ploy representing the Akaike Information Criterion as a function of each network inferred by simone is displayed.

• If "ROC" belongs to the output argument and ref.graph is specified, the ROC curve (Receiver Operating Characteristic) is plotted by representing true positive rate vs. false positive rate.

• If "PR" belongs to the output argument and ref.graph is provided by the user, the PR curve (Precision/Recall) is plotted by representing positive predicted values vs. true positive rate.

• If "path.penalty" belongs to the output argument, a regularization path is plotted by representing the value of each entry of the Theta matrix (that is, of each edge) vs. the penalty level \(\lambda\): there are as many values for the penalty as networks stocked in the simone object x.

• If "path.edges" belongs to the output argument, a regularization path is plotted by representing the value of each entry of the Theta matrix (that is, of each edge) vs. the degree of freedom in Theta (that is, the number of edges in the current network). This is done for all the network stocked in the simone object x.

• If "sequence" belongs to the output argument, an interactive plot is provided by starting from the empty network and adding the edges by successively covering the networks stocked in the simone object x.

See Also

simone.

Examples

# NOT RUN {
## data set and network generation
g    <- rNetwork(p=50, pi=50)
data <- rTranscriptData(300,g)
attach(data)

## running simone
res <- simone(X, type="steady-state")

## plotting the results: just the ROC curve
plot(res, output=c("ROC"), ref.graph=g$A)

## plotting the results: just the path as a function of the penalty
plot(res, output=c("path.penalty"), main="I want to put my own title")

## plotting the results: everything possible (the default)
plot(res)

detach(data)
# }