how to use nanmean fxn with double precision elements?

Hi, I have several variables from which I would like to calculate their mean:

ax1 ax2 ax3 ax4 ax5 are all 877x1 variables, each has many NaNs
Ax=[ax1 ax2 ax3 ax4 ax5]; %put all variables into a matrix Ax
Ax_avg=nanmean(Ax,2);   %according to a search of the documentation mean(A,dim) and nanmean(A,dim) should return the mean values for elements along the dimension specified.

However, when I run the above code I get the following error: Undefined function 'nanmean' for input arguments of type 'double'.

I assume this is saying that each element in Ax is double precision...if so, I still don't see why the mean cannot be calculated this way.

This is probably a very easy fix with someone with the 'know how', but I've been at a loss; some guidance would be greatly appreciated. Thanks.

[Andrei Bobrov]

You have Statistics Toolbox?

Answer 1

[Andrei Bobrov]

t = isnan(Ax);
A1 = Ax;
A1(t) = 0;
Ax_avg = sum(A1,2)./sum(~t,2);

[Nahir]

Thank you for the help...sorry for the delay in accepting (slipped me). I did look into it and it seems that I do have the statistics tool bar so not sure why the "nanmean" function did not work. Anyway, thanks again.