Prev: ctrl+c
Next: Fit problem
From: Jos (10584) on 1 Jul 2010 05:57
"Oleg Komarov" <oleg.komarovRemove.this(a)hotmail.it> wrote in message <i0hm71$8s7$1(a)fred.mathworks.com>...
> "Antonio Trujillo-Ortiz" <atrujo(a)uabc.edu.mx> wrote in message <i0hbsv$ker$1(a)fred.mathworks.com>...
> > Hi Joerg,
> >
> > Thanks for your answer. I'm sorry. In my initial post in the resulting matrix R, I miss to type the 3rd row. So, it must be:
> >
> > R=[6 0 3;7 2 6;5 1 2;6 1 5];
> >
> > As you can see. In the initial matrix A with NaN's,
> >
> > A = [NaN 0 3;7 2 6;5 1 2;NaN NaN 5];
> >
> > I have the columns 1 & 2 with them. With the two corresponding nanmeans=6,1. Which I need to replace with its corresponding NaN's elements.
> >
> > I'm awaiting for any hint.
> >
> > Thx.
>
>
> A = [NaN 0 3;7 2 6;5 1 2;NaN NaN 5];
>
> % column nanmeans (I suppose you have the Stats TB)
> colMeans = nanmean(A);
> % Find the position of nans
> [row,col] = find(isnan(A));
> % Replace with means
> A(isnan(A)) = colMeans(col);
>
> Oleg

Alternatively:

A = [NaN 0 3;7 2 6;5 1 2;NaN NaN 5];
B = repmat(nanmean(A,1),size(A,1),1)
q = ~isnan(A)
B() = A(q)

Jos
 | 
Pages: 1
Prev: ctrl+c
Next: Fit problem