FLCNA_LQA: Esimation of mean matrix

In FeifeiXiaoUSC/FLCNA: Simultaneous CNV Detection And Subclone Clustering With Single Cell Sequencing Data

Esimation of mean matrix

Description

Esimation of mean matrix based on local quadratic approximation (LQA).

Usage

FLCNA_LQA(
  k,
  K1,
  index.max,
  y,
  mu.t.all,
  mu.no.penal,
  sigma.all,
  alpha,
  lambda,
  iter.num = 20,
  eps.LQA = 1e-05,
  eps.diff = 1e-05
)

Arguments

k	k-th cluster index.
K1	K1 cluster in total.
index.max	Initial cluster index assigned accroding to posterior probability.
y	n by p data matrix.
mu.t.all	K by p mean matrix from previous EM-step.
mu.no.penal	K by p mean matrix of unpenalized estimates (lambda=0).
sigma.all	p by p diagnal covariance matrix.
alpha	n by K posterior probability matrix.
lambda	Tuning parameter.
iter.num	Max iterations in local quadratic approximation (LQA).
eps.LQA	LQA stop criterion.
eps.diff	Lower bound of mean difference.

Value

Estimated mu hat for k-th cluster with totally K1 cluster.

FeifeiXiaoUSC/FLCNA documentation built on March 29, 2025, 10:48 p.m.