dataSim: Simulate data

In ltsspca: Sparse Principal Component Based on Least Trimmed Squares

Description

the function that generates the simulation data set

Usage

dataSim(n = 200, p = 20, bLength = 4, a = c(0.9, 0.5, 0),
  SD = c(10, 5, 2), eps = 0, eta = 25, setting = "3", seed = 123,
  vc = NULL)

Arguments

n	number of observations
p	number of variables
bLength	the number of correlated variables in the first k blocks
a	numveric vector of length k+1 that contains the correlations between the variables in each block (the last block contains uncorrelated variables); by default is (0.9, 0.5, 0)
SD	numveric vector of length k+1 that contains the standard deviation of the variables in each block (the last block contains uncorrelated variables); by default is (10, 5, 2)
eps	proportion of outliers, default is 0
eta	parameter that contols the outlyingness, default is 25
setting	type of outliers: setting="1" generates the outliers which are outlying in the first two variables in the second block; setting="2" generates score outliers; setting="3" generates the orthogonal outliers which are easy to detect (the setting used in Hubert, et al (2016)); default is "3"
seed	random seed used to simulate the data
vc	controls the direction of the score outliers within the PC subspace, default is NULL

Value

a list with components

data	generated data matrix
ind	row indices of outliers
R	Correlation matrix of the data
Sigma	Covariance matrix of the data

ltsspca documentation built on Oct. 9, 2019, 5:06 p.m.