Yao, I wonder how likely it is that log transformations fit the data. More often I find that the flexibility of restricted cubic splines is needed.
One of the things that people take for granted when using log() is that they are assuming that k=0 in making the transformation log(x + k). It is often the case the the proper origin is not zero. This also impacts the problem you encountered. Besides all that, you can specify limits to predictors to nomogram() in a variety of ways as listed in the help file. Try that. Frank ----- Frank Harrell Department of Biostatistics, Vanderbilt University -- View this message in context: http://r.789695.n4.nabble.com/tansformation-of-variables-in-cph-from-rms-package-tp3325545p3325911.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.
