Only Changing NaN values of a vector

I want to access (202 elements) from vec A and check if isnan and change these but keep my normal values. Code to change all elements;

A = smooth(A, 'lowess'); %I want to change only NaNs

      keepInA = zeros(size(all_behav(:,1))); %created new vec
  for i = 1:size(A) %search 202 rows
      mem = i(:,1) == isnan; %try find isnan ---- not working (showing 202 0s)

end

I haven't made much progress, pretty bad at loops any help appreciated

Answer 1

[Elizabeth Drybrugh]

I am not sure if I mentioned this but changing to 0 does not help as I am using N-1 cross validation and in the literature it is not regarded as a 'suitable' method of solving the problem.

https://uk.mathworks.com/matlabcentral/fileexchange/67013-dummy-rep-nan-x-?focused=9163520&tab=function

I created my own function to solve it which works by creating a dummy vector of the same size, std and mean to then replace NaN with acceptable values. Posting this for anyone who comes across this question in google.

Cheers, walter for a response

:)

Answer 2

[Walter Roberson]

A( isnan(A) ) = 0;    %if you are changing them to 0.