\name{AbstractModel}
\alias{AbstractModel}
\alias{parameters}
\alias{parameters<-}
\title{An abstract class representing of the proficiency space for a group of examinees
}
\description{

  This is an abstract class that contains information about a model.  A
  model generally has parameters, contained in a
  \code{\link{ParameterObject}}, and hyperparameters, which depend on
  the implementation.  Depending on the subtype, a model has additional
  methods to support its use in a bigger structure.

  
}
\usage{
AbstractModel(...)
parameters(model)
parameters(model) <- value
}
\arguments{
  \item{\dots}{
    Initial values for the object:  these are usually population
    hyperparameters and vary with the implementation.  In particular,
    this list should generally contain a \code{param} argument providing
    the parameter and a \code{paramType} which should be a string
    containing the expected type of the parameters.
  }
  \item{model}{A model whose parameters are needed.}
  \item{value}{A parameter object containing the new parameters of the object.}
}
\details{

  This is an abstract class which is a parent of both the
  \code{\link{ProficiencyModel}} and \code{\link{EvidenceModel}}
  classes.  It exists mainly to pull together code and slots relating to
  the \code{param} slot, which should contain a
  \code{\link{ParameterObject}}.  The slot \code{paramType} should
  contain a string with the expected type of the parameter object.

  The function \code{parameters} access the parameter object associated
  with a model.  This must be a subtype of the type indicated by the
  \code{paramType} slot, or as a special case it can always be an object
  of type \code{NullParameter}.  

}
\value{
  This is an abstract class, and must be extended by other classes.

  The abstract class supplies the following slots:
  \item{param}{An object of type \code{\link{ParameterObject}} which
    contains the parameters for the model.}
  \item{paramType}{A character string which provides the name of a
    subtype of \code{\link{ParameterObject}}.  The \code{param} slot
    should be of this class.}
}
\references{
%% ~put references to the literature/web site here ~
}
\author{Russell Almond
}
\seealso{
  Methods:
  \code{parameters}, \code{\link{lpriorParam}}

  Extensions:
  \code{\link{ProficiencyModel}}, \code{\link{EvidenceModel}}, 

  Contained Classes:
  \code{\link{ParameterObject}}
  
}
\examples{

## Must be extended to be used.  See TimelessNormalPM.R for example.

}
\keyword{ distribution }
\keyword{ class }