Classif.quantile: Quantile Classifier

Description Usage Arguments Value Examples

View source: R/Classif.quantile.R

Description

Quantile classifier

Usage

1
Classif.quantile(obj, testset, testlabel, weighted = obj$weighted)

Arguments

obj

An FOBJ for classification

testset

Must be a matrix of testset. Each row is an observation of curve. In terms of the prediction to single curve, view pre.quantile

testlabel

Testlabel, which length should be equal to the number of rows of testset

weighted

Logical. It is to define whether weight is to be used. Default the same to obj$weighted.

Value

A list containing the classification result

MCR

The total misclassification

pre

The predictive result

FP

False positive rate if only two groups

FN

False negative rate if only two groups

Examples

 1
 2
 3
 4
 5
 6
 7
 8
 9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
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))
datacheck(X, y, t)
Index_0 <- which(y == 0)
Index_1 <- which(y == 1)
 SplitPara <- 0.8 #Split parameter
 trainIndex_0 <- sample(Index_0, SplitPara * length(Index_0))
 testIndex_0 <- Index_0[-trainIndex_0]
 trainIndex_1 <- sample(Index_1, SplitPara * length(Index_1))
 testIndex_1 <- numeric()
 for(i in Index_1){
     if(i %in% trainIndex_1 ==FALSE){
        testIndex_1 <- append(testIndex_1, i)
       }
    }
 trainset <- X[c(trainIndex_0, trainIndex_1), ]
 trainlabel <- y[c(trainIndex_0, trainIndex_1)]
  testset <- X[c(testIndex_0, testIndex_1), ]
 testlabel <- y[c(testIndex_0, testIndex_1)]
 w1 <- FOBJ(trainset, trainlabel, t)
 c1 <- Classif.quantile(w1, testset, testlabel, weighted = FALSE)

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