ProxSCC: ProxSCC

In Ivis4ml/PRScc: Proximal Gradient Smoothing Sparse Convex Clustering

Description

Implementing the ProxSCC algorithm for Convex Clustering

Usage

ProxSCC(
  X = NULL,
  U = NULL,
  Lambda,
  Gamma,
  w,
  r,
  p = nrow(X),
  n = ncol(X),
  tol = 1e-05,
  maxit = 500,
  verbose = FALSE
)

Arguments

X	Data matrix to be clustered. The rows are features, and the columns are the samples
U	Initial of clustering examplers. The rows are features, and the columns are the samples. Default is NULL
Lambda	A regularization parameter for cluster number within penalty term Lambda * w[k] * |U_,i - U_,j|_2
Gamma	A regularization parameter the number of nonzero features within penalty term Gamma * r[k] * |U_k,|_2
w	A vector of nonegative weights. w[k] denotes the weight for the k-th pair of examplers.
r	The adaptive group lasso's weights for feature sparse penalty.
p	Rows of matrix X. nrow(X) (Default).
n	Columns of matrix X. ncol(X) (Default).
tol	The convergence threshold. Default 1e-5.
maxit	The maximum iterations need for algorithms. Default 500.
verbose	False as default.

Value

U

Centroid matrix of ProxSCC

iters

number of iterations

Ivis4ml/PRScc documentation built on June 4, 2020, 9:19 a.m.