Calculate trends with regression for columns in MATLAB

1 view (last 30 days)
I have a matrix with dimensions of 19 rows and 7541502 columns. I want to get the trend of each column using regression. This is possible by using MATLAB function but requires a lot of time. The written function is as follows:
Data = box1;
[r, c] = size(Data);
X = [1:r]'; Total = nan(c,3);
for j = 1:c;
Y = Data(:, j);
[b1, bint1] = regress(Y, [ones(size(Y)) X]);
Qs = [bint1(2, :), b1(2, :)];
Total = [Total; Qs];
end
Is there a way to speed up Mfile in MATLAB? Is there a way to calculate regression without using functions?
Thanks in advance.

Accepted Answer

Abdolkarim Mohammadi
Abdolkarim Mohammadi on 3 Oct 2020
Edited: Abdolkarim Mohammadi on 3 Oct 2020
You can use parfor (parallel for loop) instead of for. The more CPU cores you have, the more speed you gain.

More Answers (0)

Community Treasure Hunt

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

Start Hunting!