Problems in Constructing Two way Matrix

Dear Mr. Berend,

I want to construct "Two way Matrix". The Algorithm is like:

1)  Data of time with censoring or events
2)  Predictor variables (genes)
3)  Resample data in step 1 and 2 by WR
4)  Apply Coxph Model to resample data. Apply Surv function to each Predictor variable, individually.
5)  Give value 1 if it's P-value is less than 0.05, otherwise 0. Save it in the first Column (Column of 0's and 1's) of the Matrix.
6)  Iterate the process 1 to 5, 10 times. And make 10 Columns of 0's and 1's.

I have done the programming for it, but gives problem in Matrix.
Thanks for replying to my mail. Hope for the best.

Best Regards
Fazli Raziq

From: Berend Hasselman <bhh at xs4all.nl>
To: Fazli Raziq <raziq_fazli at yahoo.com> 
Cc: "r-help at r-project.org" <r-help at r-project.org> 
Sent: Wednesday, 19 June 2013, 9:43
Subject: Re: [R] Problems in Constructing Two way Matrix


On 19-06-2013, at 09:30, Fazli Raziq <raziq_fazli at yahoo.com> wrote:

Hello,

Where is the problem in this programming. I want two way matrix, but it gives problem.

What is the problem?
What are you expecting and what are you getting?

The program is given below:

Thanks in advance.

Regards
Fazli Raziq


rep = 2
genes = 5
pred = c()
iter  = array (dim = c(rep, pred))
pred = array(dim = c(1, genes))
m = c()
l = 1
w = 1
for(m in 1:rep) {
t = time
c = cens
tab = cbind(t, c)

# Resampling Time and Censoring
cc = sample(c, replace = TRUE)
tt = sample(t, replace = TRUE)
xx = Surv(tt, cc)

P = predic[,1:5]
n = 310
ma = matrix(P, n)
boot <- sample(seq(n), replace=TRUE)
genes5 = ma[boot, ]


for(p in 1: genes){
i = c()

for(i in w:l) {
z = genes5[,w: l]
y = coxph(xx ~ z)

}
Pval = summary(y) $ coeff[, 5]

if( any(Pval <= 0.01) ) sign <- 1 else sign <- 0
pred[1, p] = sign[1]
}

iter [m, 1:5] = pred[1: 5]

}

Running this code gives an error message:

Error: object 'cens' not found


And what is the purpose of t = time? time is a stats function.

    [[alternative HTML version deleted]]

You are asked to NOT post in html.

Berend