plotActivations {edl}

Visualize the change of connection weights between a specific outcome and all cues.

Description

Visualize the activation or the change of activation per event.

Usage

plotActivations(
  wmlist,
  cueset,
  split = "_",
  select.outcomes = NULL,
  init.value = 0,
  add.labels = TRUE,
  add = FALSE,
  ...
)

Arguments

wmlist

A list with weightmatrices, generated by RWlearning or updateWeights.

cueset

String, which contains the combination of cues for which to calculate the activations for per learning event.

split

String, separator between cues.

select.outcomes

Optional selection of outcomes to limit the number of activations that are returned. The value of NULL (default) will return all activations. Note that specified values that are not in the weightmatrices will return the initial value without error or warning. Please use getValues for returning all outcomes in the data.

init.value

Value of connection weights for non-existing connections. Typically set to 0.

add.labels

Logical: whether or not to add labels for the lines. Defaults to TRUE, see examples.

add

Logical: whether or not to add the lines to an existing plot. Defaults to FALSE (starting a new plot).

...

Optional graphical arguments, as specified in par. These parameters are forwarded to the functions emptyPlot, lines, and text.

Value

Optionally a list with label specifications is returned, which allows to plot your own labels. This may be helpful for very long labels, and for overlapping lines.

Author(s)

Jacolien van Rij

See Also

plotCueWeights, getWeightsByOutcome, getWeightsByCue

Examples

# load example data:
data(dat)

# add obligatory columns Cues, Outcomes, and Frequency:
dat$Cues <- paste("BG", dat$Shape, dat$Color, sep="_")
dat$Outcomes <- dat$Category
dat$Frequency <- dat$Frequency1
# now use createTrainingData to sample from the specified frequencies: 
train <- createTrainingData(dat)

# this training data can be used train network:
wm <- RWlearning(train)

# plot connection weights for cue = 'cat':
plotActivations(wm, cueset="BG_cat_brown")
plotActivations(wm, cueset="BG_cat")

# plot your own labels:
labels <- plotActivations(wm, cues="BG_cat", add.labels=FALSE)
legend_margin('topright', legend=labels$labels, col=labels$col, 
    lwd=1, bty='n')

# change color and select outcomes:
out <- getValues(train$Outcomes, unique=TRUE)
out <- out[! out %in% "animal"]
labels <- plotActivations(wm, cues="BG_cat", 
    select.outcome=out, add.labels=FALSE, 
	   ylim=c(-.25,1),col=alpha(1))
lab2 <- plotActivations(wm, cues="BG_cat", add.labels=FALSE, 
    select.outcomes="animal", add=TRUE, col=2, lwd=2, xpd=TRUE)
legend('topright', legend=c("animal", labels$labels), 
    col=c(lab2$col, labels$col), lwd=c(lab2$lwd, labels$lwd), 
    lty=c(lab2$lty, labels$lty), bty="n")