in a 2-way linear model with missing data the normal equations can sometimes be separated in 2 or more independent sets of equations brought about by the nature of the missing data. As far as I know this phenomenon is called 'connectedness' of the data.
Does anyone know of a test or algorithm in R that extracts the connected sets from the data (all I need is an algorithm for a 2-way classification)
greetings,
joerg
______________________________________________ [EMAIL PROTECTED] mailing list https://www.stat.math.ethz.ch/mailman/listinfo/r-help
