FOBJ: Functional Objection

In iantsuising/quickfun:

Description

FOBJ() is used to compute imformation for within FDA framework.

Usage

FOBJ(X, y, t, optns = list(smooth = FALSE, FVEthreshold = 99.99, a0 =
  0.02, M0 = 10, par.sig = 1000), weighted = FALSE)

Arguments

X	The input matrix of observed values. Rows represent repititions or samples, and columns represent observed time points.
y	The input vector of labels. ith element of this vector should correspond to the label of ith row of data matrix X.
t	An input time grid vector.
optns	A list to control computing such as smoothing, threshold and quantile range.
weighted	Whether weight for quantile loss is to be used. FALSE is default.

Value

It returns a class "FOBJ".

DataGroupby	The list group by the label of each group
y	Vector. Label of each group
t	Original time points of observation
K	Numeber. How many groups
phi	Matrix. The functional principle componants, listed by colums
scores	List. The projection scores of each group
cutPoints	Numebr. How many componants
quantlist	List. The quantile of projection scores of each group
tau	Vector. Quantile counter
weight	Matrix. Weight for classification. Default is matrix 1. The number of colums is equal to tau
optns	List. The same to the original optns
weighted	Logical. The same to the parameter "weighted"
propotion	Vector. Contains the proportion of variance of each FPC

Examples

rm(list = ls())
library(refund)
data(DTI)
X <- DTI$cca
y <- DTI$case
t <- seq(0, 1, length.out = ncol(X))
#not run
#datacheck(X, y, t)
allData <- cbind(X, y)
allData <- allData[which(DTI$visit == 1), ]
allData <- na.omit(allData)
y <- allData[, ncol(allData)]
X <- allData[, -ncol(allData)]
t <- seq(0, 1, length.out = ncol(X))
w1 <- FOBJ(X, y, t)

iantsuising/quickfun documentation built on Nov. 4, 2019, 1:52 p.m.