R help - Jun 2014 - help plsr function

Hi All,

I want to produce scores from X using $projection. When I predict, I cannot
match the predicted scores and scores using x%*%projection.

Below is a very simple example,

set.seed(seed=1)
y <- c(1, 1, 0, 0, 1, 1, 0, 1, 1, 0, 0, 0, 1, 1, 0, 1, 1, 0, 0, 0)
x <- matrix(runif(200),nrow=20)
data <- data.frame(cbind(y,x))
colnames(data) <-
c("out","x1","x2","x3","x4","x5","x6","x7","x8","x9","x10")
data.cpls <- plsr(out~x1+x2+x3+x4+x5+x6+x7+x8+x9+x10,2,data=data)

x.new <- matrix(runif(50),nrow=5)
x.new.centered <- x.new
for (i in 1:ncol(x.new)) {
 x.mean <- mean(x.new[,i])
 x.new.centered[,i] <- (x.new[,i]-x.mean)
}
## the predicted scores from the model
(pred <- predict(data.cpls,n.comp=1:2,newdata=x.new,type="score"))
## the predicted scores using x%*%projection
cbind(x.new.centered%*%data.cpls$projection[,1],x.new.centered%*%data.cpls$projection[,2])

Can someone please tell me why the two predicted scores don't match?

Thanks,
Annie

	[[alternative HTML version deleted]]

annie Zhang <annie.zhang2010 at gmail.com> writes:
> ## the predicted scores from the model
> (pred <-
predict(data.cpls,n.comp=1:2,newdata=x.new,type="score"))
> ## the predicted scores using x%*%projection
>
cbind(x.new.centered%*%data.cpls$projection[,1],x.new.centered%*%data.cpls$projection[,2])
>
> Can someone please tell me why the two predicted scores don't match?
If you look at the code that does the prediction:
> pls:::predict.mvr
function (object, newdata, ncomp = 1:object$ncomp, comps, type =
c("response",
    "scores"), na.action = na.pass, ...) 
{
[...]
           TT <- (newX - rep(object$Xmeans, each = nobs)) %*% 
                object$projection[, comps]

you will see that it subtracts the _old X_ coloumn means from the new X
matrix, not the _new X_ coloumn means.  So

sweep(x.new, 2, data.cpls$Xmeans, "-") %*% data.cpls$projection[,1:2]

will reproduce the values from predict().

-- 
Regards,
Bj?rn-Helge Mevik