Removing invalid results

Hi, I'm having a problem removing part of my output that is invalid. Basically I have a matrix, and certain values within that matrix are implausible, and I want to replace those values with a value that won't affect calculations. So for example my matric might be [1 2 3 4 30], but 30 might be implausible, so I want to replace it with another value that doesn't affect calculations so that, for example, if i were to average the values I would get (1+2+3+4)/4 not divided by 5. Any ideas would be appreciated. Thank you

Answer 1

[Sean de Wolski]

Replace all of your implausible values (whatever criteria you use to determine that) with nan. Then apply your function to the non nan part of your data. There are some built-in nan ops already others would be easy to write.

A = [1 2 3 4 30]
A(A>10) = nan;
idx_care = ~isnan(A);
the_mean = sum(A(idx_care))/sum(idx_care);

Answer 2

[the cyclist]

 x = [1 2 3 4 30];
 x(x>10) = NaN;

 meanx = nanmean(x); % Will ignore NaN

[Sean de Wolski]

stats toolbox required!

[the cyclist]

D'oh!