I am trying to find an elegant way to calculate a moving average. Until now I have implemented it using the usual way with loops using std::vector, but since I have to use Eigen for some other calculations I thought that it would be a good idea to use Eigen through out the program. Since I am new to Eigen it would be great to get some pointers to what ways a moving average could be implemented taking advantage of Eigen's features. Any ideas?

Your question is too vague. What is the dimensionality? Regular sampling? What about the weighting scheme? etc. Anyway, I think you will still need the main outer loop. The inner one is basically a dot product.

OK, I am looking for the most simple - the moving mean. I have looked into using

Code: Select all

vector.segment(i,n);

but can't figure out how to get the sum and divide by length of the moving mean. Hope this helps.

something like that:

for(i...)
res(i) = vector.segment(i,n).mean();

?

Exactly! Works great.