Finding the mean of subsets within cellarrays

2 次查看(过去 30 天)
Hello,
I have a panel dataset as a cellarray and one cell column represents the index of a data subset within this array (long format). Now I want to create a vector with the mean of each subset. I tried the following but this does not work completely and I would be more than happy if you could help me out here. I did that several times in STATA, but there you can handle panel data directly and specify an index variable in your dataset, obviously you have to do the work in Matlab completely by yourself. :(
Regards!
x = {'A',2.5;'A',5.0;'B',2.6}
y = unique(x(:,1))
mean = zeros(length(y),1)
for i = 1:length(y)
mean(i) = mean(x(x{:,2}==y(i)));
end

采纳的回答

Andrei Bobrov
Andrei Bobrov 2012-10-29
[a,c,c] = unique(x(:,1));
avgx = [a,num2cell(accumarray(c,cell2mat(x(:,2)),[],@mean))];

更多回答(2 个)

Jing
Jing 2012-10-29
Hi,
In your code, you can't index into a cell array using ':', and if you want to use the build-in MEAN function, you should not define a variable as 'mean', because MATLAB will treat it as a variable first. I think the following code can complete your goal.
x = {'A',2.5;'A',5.0;'B',2.6};
y = unique(x(:,1))
avgx= zeros(length(y),1);
for i = 1:length(y)
avgx(i)=mean(cell2mat(x(strcmp(y{i},x(:,1)),2)));
end

Léon
Léon 2012-10-29
编辑:Léon 2012-10-29
This is exactly what I've been searching for and the speed of that is amazing. Thank you very much! :)
May I ask a question that builds on top of that?
Is it possible to apply the function on a sub-sub-set as well? Considering we have an additional date vector:
day = (28;29;30);
can we compute the mean within the subgroup, but day wise?

标签

Community Treasure Hunt

Find the treasures in MATLAB Central and discover how the community can help you!

Start Hunting!

Translated by