Unhappy with existing ellipse fitting I present a function based on Linear Least Squares and SVD, the latter symbolically resolved and not directly used/performed for the actual fit calculation.
Attachments:
