How about something like the shortest set of examples that show the different ways that the formula, data & weights arguments can be supplied to m<-lm(...) predict(m,...) (& the pitfalls)? Part of the point of this would be to illustrate the non-standard evaluations rules for the weights arguments, and the circumstances in which arguments supplied to lm() in m<-lm(...) must still be available when calling predict(m,...).
-- Tony Plate Philippe Grosjean wrote: > Hello R Wiki Team! > > Just to mention that we have now a daily backup of the Wiki. So, it > should be safe: if you notice a problem due to a bad attack of the > server (a Wiki is always rather sensitive, since everybody can edit the > content!), just tell me, and I can restore the whole server in seconds! > > Otherwise, it should be nice to have a new challenge. The last one > (about Benford's Law) did not rise much enthousiasm, although Kevin > Wright did a nice job in solving it! > > It seems that code optimisation is favorite, although I am sure that > graph/plot challenges will have their adepts too! > > So, any idea? > > Best, > > Philippe _______________________________________________ R-sig-wiki mailing list [email protected] https://stat.ethz.ch/mailman/listinfo/r-sig-wiki
