including subject_name in variable with a loop

Ubu さんによって質問されました 2012 年 3 月 20 日

24 ビュー (過去 30 日間)

Dear all,

We are running a script over several subjects in order to get mean reaction time data per condition out of each of them independently.

We have a simple loop:

subject = {subject_1 subject_2 subject_3 ...}
for i = 1: (length(subject))
... run the script

and here we would like to save our results in a new variable whose name would be accordingly "subject_1", "subject_2" and so on.. For example:

final_subject_1 = [mean1 mean2 mean3 mean4]

We thank you very much for any suggestion!

Best,

Udiubu

0 件のコメント

サインイン to comment.

3 件の回答

リンク

回答者: Grzegorz Knor 2012 年 3 月 20 日

採用された回答

eval(['final_subject_' num2str(i) '= [mean1 mean2 mean3 mean4]'])

2 件のコメント

Ubu 2012 年 3 月 20 日

Right it works, Grzeg, but we forgot to add that our subjects have strings like "WRDF" or "RWEX" and so on..

subject = {WRDF RWEX DFGR ..}

so the num2str function is not useful here..

We just used subject_1, subject_2 for making it easier to understand (but appartently it was not..) Sorry!

How would you proceed then?

Thanks

Daniel Shub 2012 年 3 月 20 日

Conceptually it is the same thing. Just replace num2str(i) with subject{i}.

サインイン to comment.

リンク

回答者: Daniel Shub 2012 年 3 月 20 日

You can do it with eval, but then in a few days you will be back asking how to use the variables. This is a FAQ, and is generally not a good idea ...

http://matlab.wikia.com/wiki/FAQ#How_can_I_create_variables_A1.2C_A2.2C....2CA10_in_a_loop.3F

0 件のコメント

サインイン to comment.

リンク

回答者: Geoff 2012 年 3 月 20 日

This has already been solved for you, but I'll chip in.

In this case, there doesn't appear to be a good reason to use dynamic names. Why not put all your data into matrix form in the same order as the subjects?

means = zeros(length(subject), 4);
for row = 1:length(subject)
    % do your stuff...
    means(row,:) = [mean1 mean2 mean3 mean4];
end

If you're concerned about looking up a subject, make a quick helper function:

subject = { etc etc etc };  % This needs to come first.
find_subj = @(name) find( cellfun(@(s) strcmp(s,name), subject) );

Now, the means for subject_2:

means( find_subj('subject_2'), : )

This approach is far more flexible than using dynamic names. Consider this: how were you intending to plot your results?

3 件のコメント

Daniel Shub 2012 年 3 月 20 日

+1 Well said. What do you mean by dynamic names? I typically think of dynamic names as being: data.('subject_1') = [mean1, mean2];

Geoff 2012 年 3 月 20 日

Yeah I guess that's the proper term for that specific syntax. But I just meant any situation where you create a variable (or field) using a name that is set at runtime. I'm a MatLab newbie so I may be a bit loose with terminology. =)

Daniel Shub 2012 年 3 月 25 日

That is a fine use of "dynamic names", I just wanted to check to make sure I understood.

サインイン to comment.

サインインしてこの質問に回答します。

現在この質問をフォロー中です

Ubu (view profile)