RWlearningMatrix: Function implementing the Rescorla-Wagner learning.

In edl: Toolbox for Error-Driven Learning Simulations with Two-Layer Networks

Description

Function implementing the Rescorla-Wagner learning.

Usage

RWlearningMatrix(
  data,
  wm = NULL,
  alpha = 0.1,
  lambda = 1,
  beta1 = 0.1,
  beta2 = 0.1,
  progress = TRUE,
  ...
)

Arguments

data	A data frame with columns Cues and Outcomes.
wm	A weightmatrix of class matrix, or a list with weight matrices. If not provided a new weightmatrix is returned. Note that the cues and outcomes do not necessarily need to be available as cues and outcomes in the weightmatrix: if not present, they will be added.
alpha	Learning parameter (scaling both positive and negative evidence adjustments), typically set to 0.1.
lambda	Constant constraining the connection strength.
beta1	Learning parameter for positive evidence, typically set to 0.1.
beta2	Learning parameter for negative evidence, typically set to 0.1.
progress	Logical: whether or not showing a progress bar (may slow down the process).
...	Parameters for the function getValues.

Value

A weightmatrix.

Author(s)

Jacolien van Rij

See Also

RescorlaWagner, updateWeights

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
head(dat)
dim(dat)

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

# this training data can actually be used train network:
wm1 <- RWlearningMatrix(train)
# comparison with a list:
wm2 <- RWlearning(train)
length(wm2)
getWM(wm2)

# in R markdown or knitr reports the progress bar should be turned off:
wm <- RWlearningMatrix(train, progress=FALSE)

# Learning in steps is also possible:
wm <- RWlearningMatrix(train[1:20,])

train[21,c("Cues", "Outcomes")]
wm <- RWlearningMatrix(train[21,], wm=wm)

edl documentation built on Sept. 20, 2021, 9:09 a.m.