Module opendp::metrics

Various implementations of Metrics (and associated Distance).

A Metric is used to measure the distance between data. Metrics are paired with a domain on which the metric can measure distance. The distance is expressed in terms of an associated type.

Example

SymmetricDistance can be paired with a domain: VectorDomain(AtomDomain(T)). In this context, the SymmetricDistance is used to measure the distance between any two vectors of elements of type T. The SymmetricDistance has an associated distance type of u32. This means that the symmetric distance between vectors is expressed in terms of a u32.

Modules

• ffi 🔒

Structs

• AbsoluteDistance
  The absolute distance between two scalar-valued aggregates.
• ChangeOneDistance
  The smallest number of changes to make two equal-length datasets equivalent.
• DiscreteDistance
  Indicates if two elements are equal to each other.
• HammingDistance
  The number of elements that differ between two equal-length datasets.
• InsertDeleteDistance
  The smallest number of insertions or deletions to make two datasets equivalent.
• LInfDistance
  Distance between score vectors with monotonicity indicator.
• LpDistance
  The $L_p$ distance between two vector-valued aggregates.
• Parallel
  The $L^0$, $L\infty$ norms of the per-partition distances between data sets.
• PartitionDistance
  The $L^0$, $L^1$, $L\infty$ norms of the per-partition distances between data sets.
• SymmetricDistance
  The smallest number of additions or removals to make two datasets equivalent.

Type Aliases

• IntDistance
  The type that represents the distance between datasets. It is used as the associated Metric::Distance type for e.g. SymmetricDistance, InsertDeleteDistance, etc.
• L1Distance
  The $L_1$ distance between two vector-valued aggregates.
• L2Distance
  The $L_2$ distance between two vector-valued aggregates.