bayespy.inference.vmp.transformations.RotateVaryingMarkovChain¶
- class bayespy.inference.vmp.transformations.RotateVaryingMarkovChain(X, B, S, B_rotator)[source]¶
Rotation for
bayespy.nodes.SwitchingGaussianMarkovChain
Assume the following model.
Constant, unit isotropic innovation noise.
Gaussian B: (1,D) x (D,K) Gaussian S: (N,1) x (K) MC X: () x (N+1,D)
No plates for X.
Methods
__init__
(X, B, S, B_rotator)bound
(R[, logdet, inv])get_bound_terms
(R[, logdet, inv])nodes
()rotate
(R[, inv, logdet])setup
()This method should be called just before optimization.