BAS (version 1.5.5)

force.heredity.bas: Post processing function to force constraints on interaction inclusion bas BMA objects

Description

This function takes the output of a bas object and allows higher order interactions to be included only if their parent lower order interactions terms are in the model, by assigning zero prior probability, and hence posterior probability, to models that do include their respective parents.

Usage

force.heredity.bas(object, prior.prob = 0.5)

Arguments

object

a bas linear model or generalized linear model object

prior.prob

prior probability that a term is included conditional on parents being included

Value

a bas object with updated models, coefficients and summaries obtained removing all models with zero prior and posterior probabilities.

See Also

Other bas methods: BAS, bas.lm(), coef.bas(), confint.coef.bas(), confint.pred.bas(), diagnostics(), fitted.bas(), image.bas(), plot.confint.bas(), predict.basglm(), predict.bas(), summary.bas(), update.bas(), variable.names.pred.bas()

Examples

# NOT RUN {
data("chickwts")
bas.chk <- bas.lm(weight ~ feed, data = chickwts)
#  summary(bas.chk)  # 2^5 = 32 models
bas.chk.int <- force.heredity.bas(bas.chk)
#  summary(bas.chk.int)  # two models now


data(Hald)
bas.hald <- bas.lm(Y ~ .^2, data = Hald)
bas.hald.int <- force.heredity.bas(bas.hald)
image(bas.hald.int)

image(bas.hald.int)

# two-way interactions
data(ToothGrowth)
ToothGrowth$dose <- factor(ToothGrowth$dose)
levels(ToothGrowth$dose) <- c("Low", "Medium", "High")
TG.bas <- bas.lm(len ~ supp * dose, data = ToothGrowth, modelprior = uniform())
TG.bas.int <- force.heredity.bas(TG.bas)
image(TG.bas.int)
# }