Rolling Average is a very common[^1] [[Smoothing Function]], reducing [[Noise]] in data but keeping the general trend. The concept is to look at a window within your data, average what's in the window, then slide the window along the dataset until you've reached the end, producing new averages along the way. If your dataset has N elements and your window size is W, you end up with an output dataset that's `N-(W-1) x 2` elements long. ![[Pasted image 20250304163026.png]] **** # Source - self - grad school # Related [^1]: but not the _only_ smoothing function