Attention Learning COVEring map model (ALCOVE)

Share:

Description

Constructs the learning submodel of an Attention Learning COVEring map model.

Usage

1
2
3
ALCOVElearning(formula,parameters=list(eta_w=.05,eta_a=.05,r=1,q=1,spf=1),humble=TRUE,
exemplar.locations,data,subset,fixed,parStruct,
random.locations=FALSE,n.locations=10,base=NULL,ntimes,replicate=TRUE)

Arguments

formula

an object of class formula (or one that can be coerced to that class): a symbolic description of the model to be fitted. For more details of model speecification, see lm or glm.

parameters

an (optional) list with (starting) values of the parameters. If no values are supplied, defaults are used.

humble

logical. If TRUE, humble teaching signal is used.

exemplar.locations

(optional) list with exemplar node locations. See details.

data

(optional) data frame for evaluation of the formula.

subset

(optional) subset of the data.

fixed

(optional) logical vector indicating whether parameters are fixed (TRUE) or freely estimable (FALSE).

parStruct

(optional) ParStruct object. Note that if parStruct is given, the ‘fixed’ argument above will be ignored.

random.locations

If no exemplar.locations are given, should they be determined randomly? If FALSE (default), then unique values of training cues are used as exemplar locations.

n.locations

Number of randomly generated exemplar locations, if random.locations = TRUE

base

which level of the criterion variable is considered the base category? Defaults to the first level.

ntimes

an optional vector with, for each repetition in the data, the total number of trials.

replicate

are the repeated series true replications, i.e., are the model parameters considered identical for each series?

Details

ALCOVE (Kruschke, 1992) is based on the GCM model, but has a mechanism to learn the attention weights. It is formulated as an ANN.

Value

A (fitted) object of class ALCOVElearning

Author(s)

Maarten Speekenbrink

References

Kruschke, J. K. (1992). ALCOVE: An exemplar-based connectionist model of category learning. Psychological Review, 99, 22-44.

Examples

1
2
3
4
5
6
7
8
9
## open weather prediction data
data(WPT)
controls <- subset(WPT,id %in% paste("C",1:16,sep=""))
## initialize model
mod <- ALCOVElearning(y~x1+x2+x3+x4,data=controls,
  fix=list(r=TRUE,q=TRUE),ntimes=rep(200,16))
## estimate free parameters (discounting first 5 trials)
## Not run: mod <- fit(mod,discount=5)
summary(mod)