1

This is my first question here. Pls lemme know if I'm unclear...

I need to find all the unique regressed lines which are touching or near (within a given tolerance) 3 or more points for a given set of N points on a plane

A simple brute force approach is to find all 3 or more points combinations from these N points and check if the corresponding regressed lines within the input tolerance

By tolerance, I mean, average absolute distance between each point and the corresponding point on the regressed line

Brianna
  • 21
  • 2
  • Can you please explain more about what you are asking and present a piece of code? Thank you so much – El.Hum Jan 09 '19 at 08:16
  • Thanks El.Ham. I use python but haven't started coding on this yet. Need to formulate my approach first. – Brianna Jan 09 '19 at 08:44
  • 1
    I think you need to reformulate what you actually want. Mathematically there are likely to be infinitely many such lines. For example if a line passes within say half the tolerance iof three points, then a small enough change to the line parameters will mean the changed line passes within the tolerance of the three points. Practically you won;t get infinitely many lines but could get a gigantic number of them – dmuir Jan 09 '19 at 09:31
  • 1
    @dmuir, I don't think it is necessarily that bad. Theoretically you can merge similar lines into one line selected by some rule. Still there is a question how to find all such lines efficiently. I can't come up with anything better than checking all lines specified by all pairs of points and then doing the merging. Do you have any better ideas? – SergGr Jan 10 '19 at 00:20
  • Glad that you mentioned it SerGr. My response would have been the same. Pls note: my question is to find all the unique (3 or more) points which are touching or near the same regressed line (within a tolerance). – Brianna Jan 10 '19 at 02:14

0 Answers0