BMA (version 3.18.19)

predict.bic.glm: Predict function for Bayesian Model Averaging for generalized linear models.

Description

Bayesian Model Averaging (BMA) accounts for the model uncertainty inherent in the variable selection problem by averaging over the best models in the model class according to approximate posterior model probability. This function predicts the response resulting from a BMA generalized linear model from given data.

Usage

# S3 method for bic.glm
predict( object, newdata, ...)

Value

The predicted values from the BMA model for each observation in newdata.

Arguments

object

a fitted object inheriting from class bic.glm.

newdata

a data frame containing observations on variables from which the predictor variables are to be selected or constructed from a formula.

...

ignored (for compatibility with generic function).

See Also

bic.glm

Examples

if (FALSE) {
# Example 1 (Gaussian)

     library(MASS)
     data(UScrime)

     f <- formula(log(y) ~  log(M)+So+log(Ed)+log(Po1)+log(Po2)+
            log(LF)+log(M.F)+log(Pop)+log(NW)+log(U1)+log(U2)+
            log(GDP)+log(Ineq)+log(Prob)+log(Time))

     bic.glm.crimeT <- bic.glm(f, data = UScrime, 
                               glm.family = gaussian())
     predict(bic.glm.crimeT, newdata = UScrime)

     bic.glm.crimeF <- bic.glm(f, data = UScrime, 
                               glm.family = gaussian(),
                               factor.type = FALSE)
     predict(bic.glm.crimeF, newdata = UScrime)
}

if (FALSE) {
# Example 2 (binomial)

     library(MASS)
     data(birthwt)

     y <- birthwt$lo
     x <- data.frame(birthwt[,-1])
     x$race <- as.factor(x$race)
     x$ht <- (x$ht>=1)+0
     x <- x[,-9]
     x$smoke <- as.factor(x$smoke)
     x$ptl <- as.factor(x$ptl)
     x$ht  <- as.factor(x$ht)

     x$ui <- as.factor(x$ui)

     bic.glm.bwT <- bic.glm(x, y, strict = FALSE, OR = 20,
                            glm.family="binomial",  
                            factor.type=TRUE)
     predict( bic.glm.bwT, newdata = x)

     bic.glm.bwF <- bic.glm(x, y, strict = FALSE, OR = 20,
                            glm.family="binomial",  
                            factor.type=FALSE)
     predict( bic.glm.bwF, newdata = x)
}

if (FALSE) {
# Example 3 (Gaussian)

     library(MASS)
     data(anorexia)

     anorexia.formula <- formula(Postwt ~ Prewt+Treat+offset(Prewt))

     bic.glm.anorexiaF <- bic.glm( anorexia.formula, data=anorexia,
                       glm.family="gaussian", factor.type=FALSE)
     predict( bic.glm.anorexiaF, newdata=anorexia)

     bic.glm.anorexiaT <- bic.glm( anorexia.formula, data=anorexia,
                           glm.family="gaussian", factor.type=TRUE)
     predict( bic.glm.anorexiaT, newdata=anorexia)
}

if (FALSE) {
# Example 4 (Gamma)

     library(survival)
     data(cancer)

     surv.t <- veteran$time
     x <- veteran[,-c(3,4)]
     x$celltype <- factor(as.character(x$celltype))
     sel<- veteran$status == 0
     x <- x[!sel,]
     surv.t <- surv.t[!sel]

     bic.glm.vaT <- bic.glm(x, y=surv.t, 
                            glm.family=Gamma(link="inverse"),
                            factor.type=TRUE)
     predict( bic.glm.vaT, x)

     bic.glm.vaF <- bic.glm(x, y=surv.t, 
                            glm.family=Gamma(link="inverse"),
                            factor.type=FALSE)
     predict( bic.glm.vaF, x)
}

# Example 5 (poisson - Yates teeth data)

     x <- rbind.data.frame(c(0, 0, 0),
                           c(0, 1, 0),
                           c(1, 0, 0),
                           c(1, 1, 1))

     y <- c(4, 16, 1, 21)
     n <- c(1,1,1,1)

     bic.glm.yatesF <- bic.glm( x, y, glm.family=poisson(),
                               weights=n, factor.type=FALSE)

     predict( bic.glm.yatesF, x)

if (FALSE) {
# Example 6 (binomial - Venables and Ripley)

    ldose <- rep(0:5, 2)
    numdead <- c(1, 4, 9, 13, 18, 20, 0, 2, 6, 10, 12, 16)
    sex <- factor(rep(c("M", "F"), c(6, 6)))
    SF <- cbind(numdead, numalive=20-numdead) 

    budworm <- cbind.data.frame(ldose = ldose, numdead = numdead,
                                sex = sex, SF = SF)
    budworm.formula <- formula(SF ~ sex*ldose)

    bic.glm.budwormF <- bic.glm( budworm.formula, data=budworm,
                   glm.family="binomial", factor.type=FALSE)
    predict(bic.glm.budwormF, newdata=budworm)

    bic.glm.budwormT <- bic.glm( budworm.formula, data=budworm,
                      glm.family="binomial", factor.type=TRUE)
    predict(bic.glm.budwormT, newdata=budworm)
}