Hi, I have tried the (MASS) LDA example (Iris) that's given on the help pages but I don't fully understand the results. Perhaps someone could be so kind to explain it to me. On the help pages it says as return values: prior means scaling svd N call What I get as results are prior means coefficients of linear discriminants proportion of trace OK, so the first 3 appear to be the same in both cases (perhaps just a different wording) and I know I can show the other return values by something like z$call. No problem. But what is the proportion of trace? Is there a simple way to get the discriminant score or do I have to manually multiply the coefficients with the data? Cheers Roland
At 01:10 PM 2/1/2003, Roland Goecke wrote: >Hi, >Is there a simple way to get the discriminant score or do I have to >manually multiply the coefficients with the data? predict.lda will generate an object with a scores component, among other things. Try ?predict.lda
At 01:10 PM 2/1/2003, Roland Goecke wrote: >But what is the proportion of trace? > Is the proportion of variance explained by each successive discriminant.
On Sat, 1 Feb 2003, Roland Goecke wrote:> I have tried the (MASS) LDA example (Iris) that's given on the help > pages but I don't fully understand the results. Perhaps someone could be > so kind to explain it to me.`It' is explained in MASS the book, as referenced on the help page. That also includes many examples of getting the scores (or you could explore the `See Also' links). [...] Note that the help page is correct, but you are failing to appreciate that a return object and its printed representation are often different. The help page describes the first (as do all R help pages). -- Brian D. Ripley, ripley at stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UK Fax: +44 1865 272595