penalties: Compute feature-wise absolute difference penalty

In kevinwang09/CPOP: Cross Platform Omics Prediction

Compute feature-wise absolute difference penalty

Description

Compute feature-wise absolute difference penalty

Compute t-stat absolute difference penalty

Usage

colmeans_penalty(x1, x2)

p_transform(penalty)

tstat_penalty(x1, x2, y1, y2)

ks_penalty(x1, x2, statistic = TRUE)

Arguments

x1	A data matrix
x2	A data matrix
penalty	A vector of penalties to be transformed where the sum is set to the length of the vector
y1	A factor response corresponding to the columns of x1
y2	A factor response corresponding to the columns of x2
statistic	Should the KS-test statistic be returned (default) or the p-value should be returned. Logical.

Value

A vector

A vector

A vector of length matching that of x1 and x2

Examples

n = 20
p = 5
x1 = matrix(rnorm(n * p, mean = 0, sd = 1), nrow = n, ncol = p)
x2 = matrix(rnorm(n * p, mean = 1, sd = 1), nrow = n, ncol = p)
colmeans_penalty(x1, x2)
n = 20
p = 5
x1 = matrix(rnorm(n * p, mean = 0, sd = 1), nrow = n, ncol = p)
y1 = factor(rbinom(n, 1, prob = 0.5), levels = c("0", "1"))
x2 = matrix(rnorm(n * p, mean = 1, sd = 1), nrow = n, ncol = p)
y2 = factor(rbinom(n, 1, prob = 0.5), levels = c("0", "1"))
tstat_penalty(x1 = x1, x2 = x2, y1 = y1, y2 = y2)
n = 20
p = 5
x1 = matrix(rnorm(n * p, mean = 0, sd = 1), nrow = n, ncol = p)
y1 = factor(rbinom(n, 1, prob = 0.5), levels = c("0", "1"))
x2 = matrix(rnorm(n * p, mean = 1, sd = 1), nrow = n, ncol = p)
y2 = factor(rbinom(n, 1, prob = 0.5), levels = c("0", "1"))
ks_penalty(x1 = x1, x2 = x2)

kevinwang09/CPOP documentation built on April 23, 2022, 11:05 a.m.