edl (version 1.1)

activationsEvents: Calculate the activations for each learning event.

Description

Calculate the activations for each learning event. The values are returned as data frame or as a list of data frames.

Usage

activationsEvents(
  wmlist,
  data,
  split = "_",
  fun = NULL,
  return.list = FALSE,
  init.value = 0,
  normalize = FALSE
)

Arguments

wmlist

A list with weightmatrices, generated by RWlearning or updateWeights, or a single weightmatrix (matrix).

data

Data frame with columns Cues and Outcomes. Number of rows should be the same as the number of weightmatrices in wmlist.

split

String, separator between cues and/or outcomes.

fun

Function to apply to the activations for events with multiple outcomes. By default (fun=NULL) the activation values for each outcome are returned. If there are learning events with multiple outcomes, the argument return.list will be automatically set to TRUE.

return.list

Logical: whether or not the activation values are returned as list or as vector. Defaults to the value FALSE, returning a vector of activation values. But this also depends on the argument fun (see more info above).

init.value

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

normalize

Logical: whether or not the activation is normalized by dividing the total activation by the number of cues. Default is FALSE. If set to TRUE, the activation reflects the average activation per cue.

Value

Vector or list of activation values (see return.list and fun for the specific conditions, and the examples below).

See Also

getWeightsByCue, getWeightsByOutcome

Other functions for calculating activations: activationsCueSet(), activationsMatrix(), activationsOutcomes(), getActivations()

Examples

# NOT RUN {
# 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
head(dat)

# now use createTrainingData to sample from the specified frequencies: 
train <- createTrainingData(dat)
head(train)

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

# Now we calculate the activations for each event:
train$Activation <- activationsEvents(wm, train)

# With multiple outcomes per event, it is better not 
# to directly assign to a new column, as a list will 
# return. See the example below:
dat$Outcomes <- paste(dat$Shape, dat$Color, sep="_")
dat$Cues <- paste("BG", dat$Category, sep="_")
dat$Frequency <- dat$Frequency1
head(dat)
train <- createTrainingData(dat)
wm <- RWlearning(train)
# This code will elicit a warning message:
# }
# NOT RUN {
    act <- activationsEvents(wm, train)
# }
# NOT RUN {
# this code will not elicit a warning:
act <- activationsEvents(wm, train, return.list=TRUE)
head(act)
# to assign one single activation value to each event,
# we could instead apply a function, for example, by
# taking the max activation per event:
train$maxAct <- activationsEvents(wm, train, fun="max")

# }