similar to: lda problem with R

I have two specific questions regarding the output of lda function in MASS. #Question1: #========= n: sample size, p: number of variables Some articles in the literature say that LDA is singular for p > n-1. However, my experimentation with lda (default arguments) for two class problems shows collinearity for p > n-2. Does anyone know why this is the case? Does lda (MASS) use a different

Dear R-help list, apparently lda from the MASS package can be used in situations with collinear variables. It only produces a warning then but at least it defines a classification rule and produces results. However, I can't find on the help page how exactly it does this. I have a suspicion (it may look at the hyperplane containing the class means, using some kind of default/trivial

Hi all, I'm working with some data: 54 variables and a column of classes, each observation as one of a possible seven different classes: > var.can3<-lda(x=dados[,c(1:28,30:54)],grouping=dados[,55],CV=TRUE) Warning message: In lda.default(x, grouping, ...) : variables are collinear > summary(var.can3) Length Class Mode class 30000 factor numeric ### why?? I

I've got something that might be a clue to understand the strange behaviour of lda in R --gui="gnome": in R --vanilla: library(MASS) load("mod23puriflda.rda") > a <- lda.default(x=mod23puriflda[,2:5],grouping=mod23puriflda[,6],CV=F) works fine. I save lda.default as a local rda file: > milda.nognome <- lda.default >

Dear R-help list, two issues regarding lda. 1) I'm puzzled by the fact that lda's in-build cross-validation gives results different from the manual cross-validation routine that I run (of course mine may be wrong, but I don't think so). See here: library(MASS) set.seed(12345) n <- 50 p <- 10 # or p<- 200 testdata <- matrix(ncol=p,nrow=n) for (i in 1:p) testdata[,i]

Hi, a colleague ran a stepwise discriminant analysis twice in a row and got different results, suggesting some "sochasticity" in the algorithms involved. I looked at her data and found that there was a lot of collinearity, so that I reckoned that maybe "stepclass" (klaR) cannot find a clear winner when trying to include a new variable and makes a random choice. Is that true?

I'm totally confused by the following: 1. I start R simply by typing R on the shell and the following works with no problems: > library(MASS) > a <- lda(x=mod23puriflda[,2:5],grouping=mod23puriflda[,6],CV=F) > summary(a) Length Class Mode prior 23 -none- numeric counts 23 -none- numeric means 92 -none- numeric scaling 16 -none- numeric lev 23

Dear All The problem is about regularization methods in multiple regression when the independent variables are collinear. A modified regularization method with two tuning parameters l1 and l2 and their product l1*l2 (Lambda 1 and Lambda 2) such that l1 takes care of ridge property and l2 takes care of LASSO property is proposed The proposed method is given

hello list, when I use method lda of the MASS package I experience a warning: variables are collinear in: lda.default(data[train, ], classes[train]) Is there an easy way to recover from this issue within the MASS package? Or how can I tell how severe this issue is at all? I understand that I shouldn't use lda at all with collinear data and should use "quadratische" (squared?)

Hi everyone, I'm trying to vectorize an application of lda to each 2D slice of a 3D array, but am running into trouble: It seems there are quite a few 2D slices that trigger either the "variables are collinear" warning, or worse, trigger a "variable appears to be constant within groups" error and fails (i.e., ceases computation rather than skips bad slice). There are

To the purpose of fitting a 2nd order polynomial (a + b*x + c*x^2) to the chunk of signal falling in a 17 consecutive samples window I wrote the following very crude script. Since I have no previous experience of using Least Square Fit with R I would appreciate your supervision and suggestion. I guess the returned coefficients of the oolynomial are: a = -1.3191398 b = 0.1233055 c = 0.9297401

I am having a problem understanding the lda package. I have a dataset here: [,1] [,2] [,3] [1,] 2.95 6.63 0 [2,] 2.53 7.79 0 [3,] 3.57 5.65 0 [4,] 3.16 5.47 0 [5,] 2.58 4.46 1 [6,] 2.16 6.22 1 [7,] 3.27 3.52 1 If I do the following; "names(d)<-c("y","x1","x2") d$x1 = d$x1 * 100 d$x2 = d$x2 * 100 g<-lda( y ~ x1 + x2, data=d) v2

##data y1 <- matrix(c(3,1,0,1,0,1,1,0,0,0,1,0,0,0,1,1,0,1,1,1), nrow = 5, byrow = TRUE) y2 <- matrix(c (3,0,10,3,3,0,0,1,1,0,0,0,0,0,1,0,1,0,0,2,1,0,1,1,0,2,1,1,4,1), nrow = 5, byrow = TRUE) y1 <- as.data.frame(y1) y2 <- as.data.frame(y2) rownames(y1) <- rownames(y2) <- paste("site", 1:5, sep = "") colnames(y1) <-

Thanks David, Yes, I am talking about the MASS package.Thank you for pointing out that these scale the same. My question is, how do I get from the V1 data: V1 1 164.4283 2 166.2492 3 170.5232 4 156.5622 5 127.7540 6 136.7704 7 136.3436 to the other set of data: + 1 -2.3769280 + 2 -2.7049437 + 3 -3.4748309 + 4 -0.9599825 + 5 4.2293774 + 6 2.6052193 + 7 2.6820884 On Mon, Sep 28, 2009

Hello I try to fit a LDA and RDA model to the same data, which has two classes. The problem now is that the training errors of the LDA model and the training error of the RDA model with alpha=0 are not the same. In my understanding this should be the case. Am I wrong? Can someone explain what the reason for this difference could be? Here my code: LDA model: =========== % x is a dataframe tmp =

Hello, I was re-generating the moduli for SSH, and during that process I noticed that, when running the following command: ssh-keygen -M screen -O prime-tests=600 -O generator=3 -f moduli-2048-01.candidates moduli-2048-01c It does not produce any errors, only the following: ebug2: ???1467763: (4) Sophie-Germain debug2: ???1467763: generator 0 != 3 debug2: ???1467764: (4) Sophie-Germain

Hello, I was re-generating the moduli for SSH, and during that process I noticed that, when running the following command: ssh-keygen -M screen -O prime-tests=600 -O generator=3 -f moduli-2048-01.candidates moduli-2048-01c It does not produce any errors, only the following: ebug2: ???1467763: (4) Sophie-Germain debug2: ???1467763: generator 0 != 3 debug2: ???1467764: (4) Sophie-Germain

Dear R users, I’m trying to understand how correlated predictors impact the Relative Importance measure in Stochastic Boosting Trees (J. Friedman). As Friedman described “ …with single decision trees (referring to Brieman’s CART algorithm), the relative importance measure is augmented by a strategy involving surrogate splits intended to uncover the masking of influential variables by others

[to R-devel, for everyone's information; thanks a lot, Peter!] >>>>> "PD" =3D=3D Peter Dalgaard BSA <p.dalgaard@biostat.ku.dk> writes: PD> Hmmm. You may need a call to PD> #include <locale.h> PD> setlocale(LC_ALL,"") PD> (or LC_CTYPE?) before your environment variables take effect. Yes, this was `it'.

Dear R-devel list members, Ben Fairbank draw it to my attention that factanal() (in the stats package) doesn't report factor correlations for oblique rotations. Looking at the source, I see that factanal also doesn't save the factor-transformation (rotation) matrix from which these correlations can be computed. I've modified the source, attached below, so that the transformation