BMS (version 0.3.4)

as.zlm: Extract a Model from a bma Object

Description

Extracts a model out of a bma object's saved models and converts it to a zlm linear model

Usage

as.zlm(bmao, model = 1)

Arguments

bmao

A bma object, e.g. resulting from a call to bms

model

The model index, in one of the following forms: An integer, denoting the rank of the model (1 for best, 2 for second-best, ...) A numeric or logical vector of length K describing which covariates are contained in the model A hexcode character describing which covariates are contained in the model

Value

zlm

Details

A bma object stores several 'best' models it encounters (cf. argument nmodel in bms). as.zlm extracts a single model and converts it to an object of class zlm, which represents a linear model estimated under Zellner's g prior. The utility model.frame allows to transfrom a zlm model into an OLS model of class lm.

See Also

bms for creating bma objects, zlm for creating zlm objects, topmodels.bma and pmp.bma for displaying the topmodels in a bma object

Check http://bms.zeugner.eu for additional help.

Examples

data(datafls)

mm=bms(datafls[,1:6],mcmc="enumeration") # do a small BMA chain
topmodels.bma(mm)[,1:5] #display the best 5 models

m2a=as.zlm(mm,4) #extract the fourth best model
summary(m2a)

# Bayesian Model Selection:
# transform the best model into an OLS model:
lm(model.frame(as.zlm(mm)))

# extract the model only containing the 5th regressor
m2b=as.zlm(mm,c(0,0,0,0,1)) 

# extract the model only containing the 5th regressor in hexcode
print(bin2hex(c(0,0,0,0,1)))
m2c=as.zlm(mm,"01")