

   FFiitt MMuullttiinnoommiiaall LLoogg--lliinneeaarr MMooddeellss

        multinom(formula, data=sys.parent(), weights, subset, na.action,
        contrasts=NULL, Hess=F, summ=0, censored=F...)

   AArrgguummeennttss::

    formula: a formula expression as for regression models, of
             the form `response ~ predictors'. The response
             should be a factor or a matrix with K columns,
             which will be interpreted as counts for each of K
             classes.  A log-linear model is fitted, with coef-
             ficients zero for the first class. An offset can
             be included: it should be a matrix with K columns
             if the response is a matrix with K columns or a
             factor with K > 2 classes, or a vector for a fac-
             tor with 2 levels.  See the documentation of `for-
             mula' for other details.

       data: an optional data frame in which to interpret the
             variables occurring in `formula'.

    weights: optional case weights in fitting.

     subset: expression saying which subset of the rows of the
             data should  be used in the fit. All observations
             are included by default.

   na.action: a function to filter missing data.

   contrasts: a list of contrasts to be used for some or all of
             the factors appearing as variables in the model
             formula.

       Hess: logical for whether the Hessian (the observed
             information matrix) should be returned.

       summ: integer; if non-zero summarize by deleting dupli-
             cate rows and adjust weights.  Methods 1 and 2
             differ in speed (2 uses `C'); method 3 also com-
             bines rows with the same X and different Y, which
             changes the baseline for the deviance.

   censored: If Y is a matrix with `K > 2' columns, interpret
             the entries as one for possible classes, zero for
             impossible classes, rather than as counts.

        ...: additional arguments for `nnet'

   DDeessccrriippttiioonn::

        Fits multinomial log-linear models via neural networks.

   VVaalluuee::

        A `nnet' object with additional structure.

   deviance: the residual deviance.

        edf: the (effective) number of degrees of freedom used
             by the model

        AIC: the AIC for this fit.

    Hessian: (if `Hess' is true).

   EExxaammpplleess::

        options(contrasts = c("contr.treatment", "contr.poly"))
        bwt.mu <- multinom(low ~ ., bwt)
        bwt.mu
        Call:
        multinom(formula = low ~ ., data = bwt)

        Coefficients:
        (Intercept)         age         lwt   raceblack   raceother
           0.823201   -0.037238   -0.015654    1.192404    0.740656
              smoke         ptd          ht          ui        ftv1
           0.755505    1.343759    1.913201    0.680202   -0.436385
              ftv2+
           0.179004

        Residual Deviance: 195.48
        AIC: 217.48

