Hello. I've got a dataset that may have outliers in both x and y. While I am not at all familiar with robust regression, it looked like the function lmrob in package robustbase should handle this situation. When I try to use it, I get:
Too many singular resamples Aborting fast_s_w_mem() Looking into it further, it appears that for an indicator variable in one of my interaction terms, 98% of the data have value 1 and only 2% have value 0. I believe this is the cause of the problem, but am confused as to why the algorithm cannot handle this situation. The probability of actually getting a singular sample ought to be fairly low, unless the sample sizes are fairly tiny. Is there some parameter I can tweak to increase the sample size, or is something else going on? You can easily reproduce this by running the following. Any advice would be appreciated. Thank you. library(robustbase) x <- rnorm(10000) isZ <- c(rep(1,9800),rep(0,200)) y <- rnorm(10000) model <- lmrob(y~x*isZ) -- View this message in context: http://r.789695.n4.nabble.com/Robust-regression-error-Too-many-singular-resamples-tp2286585p2286585.html Sent from the R help mailing list archive at Nabble.com. ______________________________________________ R-help@r-project.org mailing list https://stat.ethz.ch/mailman/listinfo/r-help PLEASE do read the posting guide http://www.R-project.org/posting-guide.html and provide commented, minimal, self-contained, reproducible code.
