assoc {logmult}

Identifying Scores from Association Models

Description

Identify log-multiplicative association scores from over-parameterized gnm models.

Usage

## S3 method for class 'rc'
assoc(model, weighting = c("marginal", "uniform", "none"),
                   rowsup = NULL, colsup = NULL, ...)

## S3 method for class 'rc.symm'
assoc(model, weighting = c("marginal", "uniform", "none"),
                        rowsup = NULL, colsup = NULL, ...)

## S3 method for class 'hmskew'
assoc(model, weighting = c("marginal", "uniform", "none"),
                       rowsup = NULL, colsup = NULL, ...)

## S3 method for class 'yrcskew'
assoc(model, weighting = c("marginal", "uniform", "none"), ...)

## S3 method for class 'rcL'
assoc(model, weighting = c("marginal", "uniform", "none"), ...)

## S3 method for class 'rcL.symm'
assoc(model, weighting = c("marginal", "uniform", "none"), ...)

## S3 method for class 'rcL.trans'
assoc(model, weighting = c("marginal", "uniform", "none"), ...)

## S3 method for class 'hmskewL'
assoc(model, weighting = c("marginal", "uniform", "none"), ...)

## S3 method for class 'rcL.trans.symm'
assoc(model, weighting = c("marginal", "uniform", "none"), ...)

Arguments

model

a gnm object, usually obtained using rc, hmskew, yrcskew, rcL, or rcL.trans, but not necessarily.

weighting

the weights to be used when normalizing scores (see ‘Details’).

rowsup

a matrix with the same columns as the model data giving supplementary (passive) rows to include in the result.

colsup

a matrix with the same rows as the model data giving supplementary (passive) columns to include in the result.

...

currently unused.

Details

These functions extract parameters from gnm log-multiplicative models and make them identifiable by imposing the required constraints on them. The general pattern is that row and column scores are separately centered around 0 and scaled so that they sum to 1, and so that their cross-dimensional correlation is null. From this operation result two series of scores (rows and columns) plus an intrinsic association coefficient (phi) for each dimension.

Most users do not need to call these directly, but they are still made public since they may be useful for advanced uses, notably when combining log-multiplicative association components with other model specifications. assoc can be used to identify the scores, the rest of the coefficients being extracted manually by the caller.

Value

An assoc object with the following components:

phi

The intrisic association parameters, one per dimension.

row

Row scores, normalized so that their (weighted) sum is 0, their (weighted) sum of squares is 1, and their (weighted) cross-dimensional correlation is null.

col

Column scores, normalized so that their (weighted) sum is 0, their (weighted) sum of squares is 1, and their (weighted) cross-dimensional correlation is null.

weighting

The name of the weighting method used, reflected by row.weights and col.weights.

row.weights

The row weights used for the identification of scores, as specified by the weighting argument.

col.weights

The column weights used for the identification of scores, as specified by the weighting argument.

See Also

rc, hmskew, yrcskew, rcL, rcL.trans