How can I repeat an iteration after if has been verified?

My current script can remove 1 outlier per row of data by performing a grubbs test and removing the maximum if it fails.

I would like the script to be able to rerun a row after it has removed an outlier, to determine if there are more outliers (and remove them if necessary). The Grubbs test used here returns a value 'h' of 0 or 1 for either not finding, or having found an outlier. All outliers are positive in this case.

The script currently replaces all '8's in matrix r, but not the secondary outlier '7.9'.

r = [8 5 5 5 5 5 7.9 5 5 5;5 8 5 5 5 5 5 5 5 5;5 5 8 5 5 5 5 5 5 5;5 5 5 8 5 5 5 5 5 5;5 5 5 5 8 5 5 5 5 5;5 5 5 5 5 8 5 5 5 5;5 5 5 5 5 5 8 5 5 5;5 5 5 5 5 5 5 8 5 5;5 5 5 5 5 5 5 5 8 5;5 5 5 5 5 5 5 5 5 8;];

h = zeros(10,1);

for g = 1:10
h(g,:) = Grubbs(r(g,:),0.05,'right');
    if h(g,:) == 1
         [value, location] = max(r(g,:));
         r(g,location) = NaN;
end

Thanks!

Answer 1

[Azzi Abdelmalek]

Use a while loop instead a for loop

g=1;
while g <10
h(g,:) = Grubbs(r(g,:),0.05,'right');
  if h(g,:) == 1
       [value, location] = max(r(g,:));
       r(g,location) = NaN;
  else
      g=g+1;
  end
end

[Pieter Kooijman]

Tnx for the reply, it seems like a good solution, but now the script does not need to rerun the Grubbs test. So 'h' will remain 1 and the if commands will repeat infinitely. Any ideas?

[Azzi Abdelmalek]

Try this

g=1;
while g <10
h(g,:) = Grubbs(r(g,:),0.05,'right');
if h(g,:) == 1
     [value, location] = max(r(g,:));
     r(g,location) = NaN;
     if all(isnan(r(g,:)))
        g=g+1;
     end 
else
    g=g+1;
end
end