Problem with the mean function

Hi everyone,

I've got a little pb with the 'mean' function :

My matrix has a class 'double' on my workspace : &lt70080*1 double> I would like to do an average of those values :

moyenne = mean(refoulement(1:8760,1)) and the answer is NaN. The values are positives.

Do you have an idea ?

Thanks !

Answer 1

[Kye Taylor]

Try this

meanWithoutNans = mean(refoulement(~isnan(refoulement)));

[Alexandre]

Ok it works perfectly. If i want to select some rows in my matrix, where should I put it ?

Here for ex : meanWithoutNans = mean(refoulement(~isnan(refoulement(1:i,:)))); ?

Ty

[Kye Taylor]

I interpret you question as, "How do I take the mean of rows in a matrix and avoid the nans?"

To do that, try

meanOfRowsWithoutNans = zeros(size(refoulement,1),1);

for i = 1:size(refoulement,1) % for each row

meanOfRowsWithoutNans(i) = ...

mean(refoulement(i,~isnan(refoulement(i,:))));

end

Answer 2

[Image Analyst]

What does this say:

hasNans = max(isnan(refoulement))

I want to see if there are any nan's in your original matrix.

Answer 3

[Sean de Wolski]

If you have the Stats Toolbox, take a peak at:

doc nanmean

Also this FEX submission:

http://www.mathworks.com/matlabcentral/fileexchange/6837-nan-suite

Answer 4

[Alexandre]

The answer for hasNans = max(isnan(refoulement)) is one thus I guess there are some NaN in the column.

What should I do then ?

I saw some topic explaining that in that case I have to code a function to get rid of the NaN, is that correct ?

Unfortunately, I don't have that toolbox.

Thanks for your help !