toys.data: Toys data

In armada: A Statistical Methodology to Select Covariates in High-Dimensional Data under Dependence

Description

toys.data is a simple simulated dataset of a binary classification problem, introduced by Weston et.al..

Usage

toys.data

Format

An object of class list of length 2.

Details

• $Y: output variable: a factor with 2 levels "-1" and "1";

• $x A data-frame containing input variables: with 30 obs. of 50 variables.

The data-frame x is composed by 2 independant clusters, each cluster contains 25 correlated variables. It is an equiprobable two class problem, Y belongs to -1,1, with 12 true variables (6 true variables in each cluster), the others being noise. The simulation model is defined through the conditional distribution of the X^j for Y=y. In the first cluster, the X^j are simulated in the following way:

• with probability 0.7, X^j ~N(y,2) for j=1,2,3, and X^j ~ N(0,2) for j=4,5,6 ;

• with probability 0.3, X^j ~ N(0,2) for j=1,2,3, and X^j ~ N(y(j-3),2) for j=4,5,6 ;

• the other variables are noise, X^j ~ N(0,1) for j=7,. . . ,25.

The second cluster of 25 variables is simulated in a similar way.

Source

Weston, J., Elisseff, A., Schoelkopf, B., Tipping, M. (2003), Use of the zero norm with linear models and Kernel methods, J. Machine Learn. Res. 3, 1439-14611

Examples

library(ClustOfVar)
library(impute)
library(FAMT)
library(VSURF)
library(glmnet)
library(anapuce)
library(qvalue)
X<-toys.data$x
Y<-toys.data$Y
scoreX<-data.frame(c(rep(8,6),rep(0,19),rep(8,6),rep(0,19)))
rownames(scoreX)<-colnames(X)
select<-ARMADA.heatmap(X, Y,  scoreX, threshold=1)
 ## Not run: 
result<-ARMADA(X,Y, nclust=2)
select<-ARMADA.heatmap(X, Y,  result[[3]], threshold=5)

## End(Not run)

armada documentation built on May 2, 2019, 6:37 a.m.