Re: Set Repeatability

ravenous.wolves_at_gmail.com
Date: 02/09/05 

Date: 9 Feb 2005 07:13:21 -0800

It's possible for points in the same run to be less than the proximity
tolerance apart. The closest two points in the same run could be would
be 5um, with a standard proximity tolerance being 10um, but anywhere
from 10-50 are workable values.

Currently, if more than 1 point from a single run is within the
tolerance of one point in the master list (and only one point), then
the closest point is chosen to match with the one from the master list,
and the others are added to the master list as unique points.

Quantizing to a fixed grid introduces the problem where points may be
very close in terms of proximity, but fall into adjacent cells.

I don't know how to handle (in a meaningful way) points from a single
run that are "the same" either based on the proximity tolerance
threshold or a fixed grid framework. These should be a small percentage
of the population in general, but I don't have a handle on how frequent
these cases are appearing in my current data sets.

Thanks,
-ken

Relevant Pages

  • Re: ratio approximation algorithm
    ... I'm trying to find the most effective algorithm for solving a certain ... What happens here is that we are given a value for X, a ratio which may ... we are given a tolerance which varies. ... for each b,c pair do a binary search to find closest value to x.b.d ...
    (comp.programming)
  • access form seach: how to set tolerances to guide results given?
    ... I want a form to give results from a database that are closest to the values ... entered (within a tolerance that I would like to be able to set). ... Prev by Date: ...
    (microsoft.public.access.forms)