

   FFiitt RRoobbuusstt LLiinneeaarr RReeggrreessssiioonn MMooddeell

        rlm(formula, data, weights, subset, na.action,
             model=F, k=1.345, sw=1000, ...)

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

    formula: a formula object, with the response on the left of
             a `~' operator, and the terms, separated by `+'
             operators, on the right.

       data: an optional `data.frame' in which to interpret the
             variables named in the formula, or in the `subset'
             and the `weights' argument.

    weights: optional weights; if supplied, the algorithm fits
             to minimize the sum of the weights multiplied into
             the squared residuals.  The weights must be
             strictly positive.

     subset: optional expression saying that only a subset of
             the rows of the data should be used in the fit.

   na.action: a missing-data filter function, applied to the
             `model.frame', after any subset argument has been
             used.

      model: flag to control what is returned.  If this is
             `TRUE', then the model frame is returned.  `X' and
             `y' are always returned.

          k: The control value for Winsorizing. The default
             gives 95% efficiency at the normal.

         sw: switch to Huber proposal 2 scale at iteration `sw'
             and beyond.

        ...: additional arguments for the fitting routines.
             The most likely one is `maxit', which sets the
             iteration limit, by default 20.

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

        Fits a robust linear regression model, using an M-esti-
        mator with Huber's psi function.

   DDeettaaiillss::

        The fit uses Huber's M-estimator, and initially uses
        the median absolute deviation scale estimate based on
        the residuals. This can be changed to Huber's proposal
        2 after `sw' iterations.

        Generic functions such as `print' and `summary' have
        methods to show the results of the fit.

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

        an object of class `rlm' representing the fit, inherit-
        ing from `lm'.  This has all the components of an `lm'
        object, plus `k,' the scale `s' and `conv' which is a
        vector monitoring the convergence.

   SSeeee AAllssoo::

        `summary.rlm'

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

        data(phones)
        attach(phones)
        res <- rlm(calls ~ year)
        print(res)

        data(stack)
        rlm(stack.loss ~ stack.x)

