trilearn_intra-class.rst

trilearn_intra-class

Graph intra-class

Package	trilearn
Version	1.25
Language	Python
Docs
Paper
Graph type	UG
Docker	bpimages/trilearn:2.0.5
Module	trilearn_intra-class

Description

An object of the intraclass module defines a zero mean multivariate Gaussian distribution by its covariance matrix \Sigma as

\Sigma_{ij} = \begin{cases}
    \sigma^2, &\text{ if } i=j\\
    \rho\sigma^2, &\text{ if } (i,j) \in E \\
\end{cases}

and \Sigma^{-1}_{ij} = 0 \text{ if } (i, j) \in E \text{, where } \sigma^2 > 0 \text{ and } \rho \in [0, 1] denote the variance and correlation coefficient, respectively. Using an object id of this module in the parameters_id field of the data section requires that graph_id represents a decomposable graph.

Example

[
  {
    "id": "intra-class",
    "rho": 0.4,
    "sigma2": 1.0
  }
]

.. footbibliography::