gendata_FAM | R Documentation |
This function generates inputs X
given by the response variable y
using a multivariate normal model.
gendata_FAM(n, muj, A, sd_g = 0, stdx = FALSE)
n |
Number of observations. |
muj |
C by p matrix, with row c representing y = c, and column j representing x_j.
Used to specify |
A |
Factor loading matrix of size p by p, see details. |
sd_g |
Numeric value indicating noise level δ, see details. |
stdx |
Logical; if |
The means of each covariate x_j depend on y
specified by the
matrix muj
; the covariate matrix Σ of the multivariate normal
is equal to AA^tδ^2I, where A
is the factor loading matrix
and δ is the noise level.
A list contains input matrix X
, response variables y
,
covariate matrix SGM
and muj
(standardized if stdx = TRUE
).
gendata_MLR
## feature #1: marginally related feature ## feature #2: marginally unrelated feature, but feature #2 is correlated with feature #1 ## feature #3-5: marginally related features and also internally correlated ## feature #6-10: noise features without relationship with the y set.seed(12345) n <- 100 p <- 10 means <- rbind( c(0, 1, 0), c(0, 0, 0), c(0, 0, 1), c(0, 0, 1), c(0, 0, 1) ) * 2 means <- rbind(means, matrix(0, p - 5, 3)) A <- diag(1, p) A[1:5, 1:3] <- rbind( c(1, 0, 0), c(2, 1, 0), c(0, 0, 1), c(0, 0, 1), c(0, 0, 1) ) dat <- gendata_FAM(n, means, A, sd_g = 0.5, stdx = TRUE) ggplot2::qplot(dat$y, bins = 6) corrplot::corrplot(cor(dat$X))
Add the following code to your website.
For more information on customizing the embed code, read Embedding Snippets.