Rev Language Reference

model(RevObject x, RevObject ...)

x :	RevObject (pass by const reference)
	Any variable that is connected in the model graph.
... :	RevObject (pass by const reference)
	Additional variables.

`model(x)` creates a model object by creating a copy of all elements and parameters that influence or are influenced by the likelihood of `x`. Because `model` works with copies of objects, conducting an mcmc(mc) analysis on a model object will not change the values of the objects in the RevBayes workspace. The model object can be modified to ignore specific data elements using the method `ignoreData`. Thus to run without the sequence data `phySeq` you might specify: mymodel.ignoreData(phySeq) Only clamped nodes can be ignored. To ignore all clamped nodes you can use the method `ignoreAllData`: mymodel.ignoreAllData()

# Create a simple model (unclamped)
a ~ dnExponential(1)
b ~ dnExponential(a)
mymodel = model(b) # equivalent to model(a) or model(a, b)

# Save a DOT visualization of the model to file
mymodel.graph("mymodel.dot")

# Create a move vector and a monitor vector
moves = [ mvScale( a, lambda = 1.0, weight = 1.0 ) ]
monitors = [ mnScreen(printgen = 10, a) ]

# Create an mcmc object
mymcmcObject = mcmc( mymodel, monitors, moves )

# Print value of a
print(a)

# Run a short analysis
mymcmcObject.run( generations = 100 )

print(a) # Value is unchanged in the workspace - only the copy is modified

• graph(String file, Bool verbose, String bg)
• ignoreAllData()
• ignoreData(RevObject ...)