standardization
Peter Dalgaard wrote:
SAS Enterprise Miner recommendeds to standardize using X / STDEV(X) versus [X ? mean(X)] / STDEV(X)
This makes absolutely no sense out of context.
To paraphrase Tanenbaum: "The nice thing about standardization is that there's so many ways to do it". Baz [[ Free On-line Dictionary of Computing: Andrew Tanenbaum, in his Computer Networks book, once said, "The nice thing about standards is that there are so many of them to choose from", ]]