The hardware and bandwidth for this mirror is donated by dogado GmbH, the Webhosting and Full Service-Cloud Provider. Check out our Wordpress Tutorial.
If you wish to report a bug, or if you are interested in having us mirror your free-software or open-source project, please feel free to contact us at mirror[@]dogado.de.

Metodiev Martin

2026-09-14

Preparing the data

Load the galaxy dataset:

![plot of chunk load data](figure/load data-1.png)

Computing the sample from MCMC output using bayesmix

Define the logposterior as a function of G and y:

Calculating the MCMC output for 2 components:

## 
## Attaching package: 'bayesmix'

## The following object is masked from 'package:Rfast':
## 
##     Sort

## Compiling model graph
##    Declaring variables
##    Resolving undeclared variables
##    Allocating nodes
## Graph information:
##    Observed stochastic nodes: 82
##    Unobserved stochastic nodes: 88
##    Total graph size: 345
## 
## Initializing model

![plot of chunk mcmc output for 2 components](figure/mcmc output for 2 components-1.png)![plot of chunk mcmc output for 2 components](figure/mcmc output for 2 components-2.png)![plot of chunk mcmc output for 2 components](figure/mcmc output for 2 components-3.png)

Same for 3 components:

## Compiling model graph
##    Declaring variables
##    Resolving undeclared variables
##    Allocating nodes
## Graph information:
##    Observed stochastic nodes: 82
##    Unobserved stochastic nodes: 90
##    Total graph size: 348
## 
## Initializing model

![plot of chunk mcmc output for 3 components](figure/mcmc output for 3 components-1.png)![plot of chunk mcmc output for 3 components](figure/mcmc output for 3 components-2.png)![plot of chunk mcmc output for 3 components](figure/mcmc output for 3 components-3.png)

Computing the THAMES

Estimates of the logarithm of the marginal likelihood:

## radius is too large, decreasing radius ...
## radius is too large, decreasing radius ...

##      G = 2     G = 3
## 1 -235.254 -226.6771

The marginal likelihood is maximised at G=3, which indicates evidence for the Gaussian mixture model with G=3 components. It is also possible to plot the overlap graphs and compute the criterion of overlap (co):

![plot of chunk plot overlap g2 and g3](figure/plot overlap g2 and g3-1.png)![plot of chunk plot overlap g2 and g3](figure/plot overlap g2 and g3-2.png)

##   G = 2 G = 3
## 1     2     3

The CO is largest for G=3, which indicates that the mixture components are most distringuishable when G=3. The CO can also be computed directly, without computing the THAMES:

![plot of chunk plot overlap g2 and g3 using overlapgraph function](figure/plot overlap g2 and g3 using overlapgraph function-1.png)![plot of chunk plot overlap g2 and g3 using overlapgraph function](figure/plot overlap g2 and g3 using overlapgraph function-2.png)

##   G = 2 G = 3
## 1     2     3

These binaries (installable software) and packages are in development.
They may not be fully stable and should be used with caution. We make no claims about them.
Health stats visible at Monitor.