superpc.predict {superpc}R Documentation

Form principal components predictor from a trained superpc object

Description

Computes supervised principal components, using scores from "object"

Usage

superpc.predict(object, data, newdata, threshold, n.components = 3, prediction.type = c("continuous", "discrete", "nonzero"), n.class = 2)

Arguments

object Obect returned by superpc.train
data List of training data, of form described in superpc.train documentation,
newdata List of test data; same form as training data
threshold Threshold for scores: features with abs(score)>threshold are retained.
n.components Number of principal components to compute. Should be 1,2 or 3.
prediction.type "continuous" for raw principal component(s); "discrete" for principal component categorized in equal bins; "nonzero" for indices of features that pass the threshold
n.class Number of classes into which predictor is binned (for prediction.type="discrete"

Value

v.pred Supervised principal componients predictor
u U matrix from svd of feature matrix x
d singual values from svd of feature matrix x
which.features Indices of features exceeding threshold
n.components Number of supervised principal components requested
call calling sequence

Author(s)

Eric Bair and Robert Tibshirani

Examples

set.seed(332)
#generate some data

x<-matrix(rnorm(1000*20),ncol=20)
y<-10+svd(x[1:30,])$v[,1]+ .1*rnorm(20)
ytest<-10+svd(x[1:30,])$v[,1]+ .1*rnorm(20)
censoring.status<- sample(c(rep(1,17),rep(0,3)))
censoring.status.test<- sample(c(rep(1,17),rep(0,3)))

featurenames <- paste("feature",as.character(1:1000),sep="")
data<-list(x=x,y=y, censoring.status=censoring.status, featurenames=featurenames)
data.test<-list(x=x,y=ytest, censoring.status=censoring.status.test, featurenames= featurenames)

a<- superpc.train(data, type="survival")

fit<- superpc.predict(a, data, data.test, threshold=1.0, n.components=1)

plot(fit$v.pred,ytest)


[Package superpc version 1.05 Index]