RSiena (version 1.1-212)

iwlsm: Function to fit an iterated weighted least squares model.

Description

Fits an iterated weighted least squares model.

Usage

iwlsm(x, ...)
## S3 method for class 'formula':
iwlsm(formula, data, weights, ses, \dots, subset, na.action,
    method = c("M", "MM", "model.frame"),
    wt.method = c("inv.var", "case"),
    model = TRUE, x.ret = TRUE, y.ret = FALSE, contrasts = NULL)
## S3 method for class 'default':
iwlsm(x, y, weights, ses, \dots, w = rep(1/nrow(x), nrow(x)),
    init = "ls", psi = psi.iwlsm,
    scale.est = c("MAD", "Huber", "proposal 2"), k2 = 1.345,
    method = c("M", "MM"), wt.method = c("inv.var", "case"),
    maxit = 20, acc = 1e-4, test.vec = "resid", lqs.control = NULL)
psi.iwlsm(u, k, deriv = 0, w, sj2, hh)

Value

An object of class "iwlsm" inheriting from "lm". Note that the df.residual component is deliberately set to NA to avoid inappropriate estimation of the residual scale from the residual mean square by "lm" methods.

The additional components not in an lm object aresthe robust scale estimate usedwthe weights used in the IWLS processpsithe psi function with parameters substitutedconvthe convergence criteria at each iterationconvergeddid the IWLS converge?wresida working residual, weighted for "inv.var" weights only.

Details

This function is very slightly adapted from rlm in packages MASS. It alternates between weighted least squares and estimation of variance on the basis of a common variance. The function psi.iwlsm calculates the weights for the next iteration. Used by siena08 to combine estimates from different sienaFits.

References

Venables, W. N. and Ripley, B. D. (2002) Modern Applied Statistics with S. Fourth edition. Springer. See also http://www.stats.ox.ac.uk/~snijders/siena/

See Also

siena08, sienaMeta, sienaFit

Examples

##not enough data here for a sensible example, but shows the idea.
mymodel <- sienaModelCreate(fn=simstats0c, nsub=2, n3=100)
mynet1 <- sienaNet(array(c(s501, s502), dim=c(50, 50, 2)))
mynet2 <- sienaNet(array(c(s502, s503), dim=c(50, 50, 2)))
mydata1 <- sienaDataCreate(mynet1)
mydata2 <- sienaDataCreate(mynet2)
myeff1 <- getEffects(mydata1)
myeff2 <- getEffects(mydata2)
myeff1 <- setEffect(myeff1, transTrip, fix=TRUE, test=TRUE)
myeff2 <- setEffect(myeff2, transTrip, fix=TRUE, test=TRUE)
myeff1 <- setEffect(myeff1, cycle3, fix=TRUE, test=TRUE)
myeff2 <- setEffect(myeff2, cycle3, fix=TRUE, test=TRUE)
ans1 <- siena07(mymodel, data=mydata1, effects=myeff1, batch=TRUE)
ans2 <- siena07(mymodel, data=mydata2, effects=myeff2, batch=TRUE)
meta <- siena08(ans1, ans2)
metadf <- split(meta$thetadf, meta$thetadf$effects)[[1]]
metalm <- iwlsm(theta ~ tconv, metadf, ses=se^2)