| acts2probs {ndl} | R Documentation |
Calculate probability matrix from activation matrix, as well as predicted values
Description
acts2probs takes the activation matrix returned by
ndlClassify and calculates the matrix of probabilities
for the estimated activation matrix, as well as the predicted values
of the response variable.
Usage
acts2probs(acts)
Arguments
acts |
A matrix of activations (number of observations by number of levels of the response variable). |
Details
Probabilities in p are obtained by adding the absolute value
of the minimum activation to the activation matrix, and
renorming. The selection rule used to produce predicted is to
choose for each instance in the data the outcome value that has
received the highest probability estimate.
Value
A list with the following components:
pa matrix with the probabilities of the levels of the response variable for each observation.
predicteda character vector with predicted values.
Author(s)
Harald Baayen and Antti Arppe
References
Arppe, A. and Baayen, R. H. (in prep.). Statistical classification and principles of human learning.
See Also
See also ndlClassify.
Examples
data(think)
think.ndl <- ndlClassify(Lexeme ~ Person + Number + Agent + Register, data=think)
pdata <- acts2probs(think.ndl$activationMatrix)