Yet Another Vectorization Problem

[dirac]

Hi everyone,

This is another vectorization problem. I understand that there have been a
few of these problems on here recently and I have read them but seem to get
no furthur with my particular code...

I have the data set with 10002 rows and 21 columns. Basically I read inthe
matrix and then define other data sets/vectors from the data read in. I then
want to take a moving average (I think it's called that!) for the new vector
named data (which is 10001x20 in dimensions). What I am doing is using the
sum function for calculate a mean value for the first 1000 data points for
all 20 columns. I then want to move on to the data points between 2 and
1001; then 3 and 1002 etc...

My code reads as follows:

function av=aver(filename)

  f = dlmread(filename);        %read in the file to be analysed

  lambda = f(2:length(f),1);   %this is the 1st column from the file
  data = f(2:length(f),2:21);   %this is the 2nd to 21st columns from the
file

  av=[]   Defining the running mean as a matrix.


  for i = 1:1:9001
    av(i,:) = [(sum(data(i:999+i,:),1))/1000];  %the data(i:999+i,:) bit is
doing what I described above
  end

  end

The problem is is that it seems to run very slowly looping 9001 times and I
have never actually finished running it! And a nieve quiestion is can there
be a faster way?

Thanks in advance :)

Martin

-----
Pretty much convinced Octave>Excel for scientific data analysis!
-- 
View this message in context: 
http://octave.1599824.n4.nabble.com/Yet-Another-Vectorization-Problem-tp2294167p2294167.html
Sent from the Octave - General mailing list archive at Nabble.com.