Package 'RSSRMR'

Title: Robust Self-Representation Sparse Reconstruction and Manifold Regularization
Description: Feature selection and clustering classification under the presence of multivariate outliers in high-dimensional unlabeled data.
Authors: Abdul Wahid [aut, cre]
Maintainer: Abdul Wahid <[email protected]>
License: GPL-3
Version: 0.1.0
Built: 2026-05-29 12:56:22 UTC
Source: https://github.com/cran/RSSRMR

Help Index

Robust Distance Weights

Description

Computes robust observation weights.

Usage

dweight(x, q = 0.9)

Arguments

x

Numeric data matrix.
q

Quantile threshold.

Value

Numeric vector of weights.

Examples

set.seed(123)

x <- matrix(rnorm(50), nrow = 10)

dweight(x)

Robust Self-Representation Sparse Reconstruction and Manifold Regularization

Description

Performs robust sparse self-representation with manifold regularization.

Usage

RSSRMR(x, Wt, L, alpha = 1, beta = 1, epsilon = 0.001, maxites = 50)

Arguments

x

Numeric data matrix.
Wt

Weight matrix.
L

Laplacian matrix.
alpha

Regularization parameter.
beta

Graph regularization parameter.
epsilon

Convergence threshold.
maxites

Maximum number of iterations.

Value

A list containing:

Optimum.A

Coefficient matrix

Optimum.G

Diagonal weight matrix

Examples

set.seed(6542)

cluster1 <- matrix(
  rnorm(12 * 5, mean = 2, sd = 0.5),
  nrow = 12
)

cluster2 <- matrix(
  rnorm(13 * 5, mean = 7, sd = 0.5),
  nrow = 13
)

X <- rbind(cluster1, cluster2)

wd <- diag(runif(25))
lp <- diag(runif(25))

fit <- RSSRMR(
  x = X,
  Wt = wd,
  L = lp
)

fit$Optimum.G