mrpy.extra.reparameterise¶
Provides classes which implement variations of the MRP, in which the parameters have been transformed.
Each transformation has three available classes:
- one with a suffix MRP, which implements the core MRP quantities (i.e. is a subclass of
mrpy.core.MRP
)- one with a suffix PerObj, which extends the previous one for likelihoods based on samples of variates (i.e. is a subclass of
mrpy.likelihoods.PerObjLike
)- one with a suffix Curve, which extends the base one for likelihoods based on chi-squared minimzation against binned data (i.e. a subclass of
mrpy.likelihoods.CurveLike
)
In all, the transformed parameters are denoted p1,p2,p3
.
In addition, base classes for each are provided, which makes it easy to implement
arbitrary transformations. See the docs for ReparameteriseMRP
for more
details.
Classes
AP1Curve ([p1, p2, p3, logHs, alpha, beta]) |
|
AP1Sample ([p1, p2, p3, logHs, alpha, beta]) |
|
Ap1MRP ([p1, p2, p3, logHs, alpha, beta]) |
A fairly standard parameterisation of the TGGD (eg. |
GG2Curve ([p1, p2, p3, logHs, alpha, beta]) |
|
GG2MRP ([p1, p2, p3, logHs, alpha, beta]) |
A reparameterisation of the standard MRP form of the TGGD (eg. |
GG2Sample ([p1, p2, p3, logHs, alpha, beta]) |
|
GG3Curve ([p1, p2, p3, logHs, alpha, beta]) |
|
GG3MRP ([p1, p2, p3, logHs, alpha, beta]) |
A reparameterisation of the standard MRP form of the TGGD (eg. |
GG3Sample ([p1, p2, p3, logHs, alpha, beta]) |
|
HTCurve ([p1, p2, p3, logHs, alpha, beta]) |
|
HTMRP ([p1, p2, p3, logHs, alpha, beta]) |
A reparameterisation of the standard MRP form of the TGGD. |
HTSample ([p1, p2, p3, logHs, alpha, beta]) |
|
ReparameteriseCurveLike ([p1, p2, p3, logHs, …]) |
An extension of ReparameteriseMRP which adds necessary methods for calculating jacobians and hessians for chi-square likelihoods. |
ReparameteriseMRP ([p1, p2, p3, logHs, …]) |
Base class for reparameterising the MRP. |
ReparameteriseSampleLike ([p1, p2, p3, …]) |
An extension of ReparameteriseMRP which adds necessary methods for calculating jacobians and hessians for per-object likelihoods. |