Selection of a proper epsilon parameter depends highly on the order of magnitude of numbers included in points. I want to simplify that difficult choice introducing a percentage tolerance as second way to parametrize the series reduction algorithm. My idea is to calculate the difference between maximum and minimum value and take o given percent of it.
Selection of a proper epsilon parameter depends highly on the order of magnitude of numbers included in points. I want to simplify that difficult choice introducing a percentage tolerance as second way to parametrize the series reduction algorithm. My idea is to calculate the difference between maximum and minimum value and take o given percent of it.