The fisher transform appears easy to do but has proven difficult for me to get working with accurate precision. I have found conflicting information on it and have found some of the details to be confusing.
Specifically, someone mentions using arctan-h from numpy but I have found this to produce bizarre results that are inconsistent with the classic fisher transform indicator. If you can show me that arctan-h works accurately, then I am fine with that. It think the issue is due to the way the bounds are calculated on values at the upper and lower limits of the function.
What I need is a function that will calculate the fisher transform in a live trading environment. It needs to be able to continue calculating the transform for new data. Whatever way most effectively achieves this is what I need, but I would prefer a simple solution where a list of prices can be supplied to the function and the modified list is returned after the transform. Since it appears to be calculated iteratively, I am a little confused on the most efficient way to do this (since I don't want to have to re-calculate every value when the list rolls over with a new price)
I am an intermediate python programmer who happens to be stumped on this silly thing, and maybe you can help?
I will be testing code to make sure it is accurate and will discuss any needed changes (if any).
Submissions are private