Philosophically Incoherent Design of "Moment" Transforms
jacobnzw opened this issue · 0 comments
StateSpaceInference sub-classes are specific to a particular density (i.e. GaussianInference, StudentInference). That is, sub-classes are designed with the Bayesian viewpoint, which imposes density on the joint p(x, z). But the same philosophy isn't followed for moment transforms!
For reasons of consistency, the same Bayesian philosophy should be reflected in the class hierarchy of the "moment" transforms (they should really be referred to as statistic transforms).
The class hierarchy might look like this:
GaussianTransformorGaussianStatisticTransformGaussianSigmaPointTransform
StudentTransformorStudentStatisticTransformStudentSigmaPointTransform
There might be a common base class StatisticTransform, but currently I don't see the need for it.
Consequently this will also require redesign of the StudentInference and minor changes in GaussianInference.
The following dictionary might be helpful
| Bayesian world | Moment world |
|---|---|
| Statistic transform | Moment transform |
| Quadrature | Sigma-point rule |
| Moments | |
| Gaussian filters | Kalman filter |
| Studentian filters |
Originally posted by @jacobnzw in https://github.com/_render_node/MDEyOklzc3VlQ29tbWVudDQ0MTYwMjk2OA==/timeline/issue_comment#issuecomment-441602968