predict.spikeslab {spikeslab}

Spike and Slab Prediction

Description

Prediction on test data using spike and slab regression.

Usage

## S3 method for class 'spikeslab'
predict(object, newdata = NULL, ...)

Arguments

object

An object of class spikeslab.

newdata

Data frame or x-matrix containing test data (if omitted, the training data is used).

...

Further arguments passed to or from other methods.

Details

Computes the predicted value using a test data set.

Value

A vector of fitted values for the BMA and gnet and a matrix of fitted values for the gnet path. Also returns the grr mixing predictor if the object has been parsed by the mixing wrapper.

Author(s)

Hemant Ishwaran ([e-mail])

J. Sunil Rao ([e-mail])

Udaya B. Kogalur ([e-mail])

References

Ishwaran H. and Rao J.S. (2003). Detecting differentially expressed genes in microarrays using Bayesian model selection. J. Amer. Stat. Assoc., 98:438-455.

Ishwaran H. and Rao J.S. (2005a). Spike and slab variable selection: frequentist and Bayesian strategies. Ann. Statist., 33:730-773.

Ishwaran H. and Rao J.S. (2005b). Spike and slab gene selection for multigroup microarray data. J. Amer. Stat. Assoc., 100:764-780.

Ishwaran H. and Rao J.S. (2010). Generalized ridge regression: geometry and computational solutions when p is larger than n.

Ishwaran H., Kogalur U.B. and Rao J.S. (2010). spikeslab: prediction and variable selection using spike and slab regression. R Journal, 2(2), 68-73.

Ishwaran H. and Rao J.S. (2011). Mixing generalized ridge regressions.

See Also

spikeslab.

Examples

## Not run: 

#------------------------------------------------------------
# Example 1: get the predictor for the training data
#------------------------------------------------------------
data(diabetesI, package = "spikeslab")
x <- diabetesI[, -1]
y <- diabetesI[, 1]
obj <- spikeslab(x = x, y = y)
#gnet predictor
yhat.gnet <- predict(obj)$yhat.gnet
#an equivalent call is...
yhat.gnet <- predict(obj, x = x)$yhat.gnet

#------------------------------------------------------------
# Example 2: ozone data with interactions
#------------------------------------------------------------

data(ozoneI, package = "spikeslab")
train.pt <- sample(1:nrow(ozoneI), nrow(ozoneI) * 0.80)
obj <- spikeslab(ozone ~ . , ozoneI[train.pt, ])
ytest <- ozoneI$ozone[-train.pt]
ss.pred <- predict(obj, ozoneI[-train.pt, ])
yhat.bma <- ss.pred$yhat.bma
yhat.gnet <- ss.pred$yhat.gnet
plot(ytest, yhat.bma, ylab = "yhat", pch = 16, col = 4)
points(ytest, yhat.gnet, pch = 16, col = 2)
abline(0, 1, lty = 2, col = 2)
legend("bottomright", legend = c("bma", "gnet"), col = c(4, 2), pch = 16)

## End(Not run)