Approximation by Constraints

1
Approximation by Constraints

• Why do we need approximation/interpolation?
• Record values at some finite set of points in
  space and time.
• Does not capture the entire continuously changing
  surface.
• Need to interpolate approximation of the value
  at a given place and time from values at other
  places and time.

2
Advantage of representing approximation by
constraint databases.

• Interpolation functions is usually represented by
  constraints.
• Data compression constraint representation
  requires fewer tuples and less compute memory
  than the original representation.
• Queries can often be evaluated much faster on the
  constraint approximation.

3
Triangulated Irregular Networks

• Often used in GIS to approximate the surface of a
  landscape.
• Landscape surfaces
• Hard to represent in database systems infinite
  number of points.
• TIN approximation
• take a sample of points on the surface
• find a triangulation of the sample points