Generating a for loop and writing data

I have a lot of data and need to do many stepwise regressions. The data is set out in 3 columns and 12040 rows. Each regression should include only 8 rows of data e.g. regression 1 1:8, regression 2 9:16 etc

So far the important part of my code looks like this (this gives me the correct output for the first regression)

xx=data(1:8,2:3); yy=data(1:8,1); [B,SE,PVAL,INMODEL,STATS,NEXTSTEP,HISTORY]=stepwisefit(xx,yy,'penter',.05);

I need to put this in a loop so that it will cycle through as specified above, what should this loop look like?

As this will spit out a lot of data, is there a way for the Coefficients and P values to be written to a file each time?

Any help is greatly appreciated as I'm relatively new to matlab :)

Answer 1

[Shoaibur Rahman]

for i = 1:size(data,1)/8
  xx=data(8*i-7:8*i,2:3); yy=data(8*i-7:8*i,1);
  % append other stuffs here, use loop index to save B, SE, PVAL etc. values.
end

Mare sure that the data length (number of rows) is multiple of 8. If not, you can use ceil, floor, round, etc. functions, or append some rows of zeros at the end each column as per your requirement. I hope this should help.

[Henrik]

Damn, you beat me to it!

One comment, though, it's not always a good idea to use i as index, since MATLAB also uses it for the imaginary unit.

[Shoaibur Rahman]

Hi Henrik,

Thanks for that note. Just to share, this is true both for i and j. So, it is better not to use those as vector or matrix indices. However, if I am certain that I am not handing complex numbers in any part of the code, then it is okay. Yet, good practice is to avoid those. Thanks.

Answer 2

[Henrik]

This could be a start. I don't know exactly what output stepwisefit gives, and which parts of it you want to save, but I've given some examples of how to store B that hopefully will lead you the right way.

sz=size(data);
for k=1:sz(1)/8
xx=data((k-1)*8+(1:8),2:3);
yy=data((k-1)*8+(1:8),1);
[B,SE,PVAL,INMODEL,STATS,NEXTSTEP,HISTORY]=stepwisefit(xx,yy,'penter',.05);
%examples of how to save B, depending on what you want
%[B(k),SE,PVAL,INMODEL,STATS,NEXTSTEP,HISTORY]=stepwisefit(xx,yy,'penter',.05); 
%[B(k,:),SE,PVAL,INMODEL,STATS,NEXTSTEP,HISTORY]=stepwisefit(xx,yy,'penter',.05);
%[B{k},SE,PVAL,INMODEL,STATS,NEXTSTEP,HISTORY]=stepwisefit(xx,yy,'penter',.05);
end