cbird: Clustering of multivariate binary data with dimension...

In cbird: Clustering of Multivariate Binary Data with Dimension Reduction via L1-Regularized Likelihood Maximization

Description

This function conducts the clustering of binary data with reducing the dimensionality (CLUSBIRD) proposed by Yamamoto and Hayashi (2015).

Usage

cbird(Y, N.comp, N.clust, lambda=0, N.ite=10000, N.random=1,
     show.random.ite=FALSE, eps=0.0001, mc.cores=1)

Arguments

Y	Binary data matrix (N * D), where N denotes sample size and D denotes the number of binary variables (0 or 1).
N.comp	The number of component to be extracted.
N.clust	The number of mixture components, which corresponds to the number of clusters.
lambda	A tuning parameter of an L1 penalty for loadings. A non-negative real value should be used as the value of lambda.
N.ite	The number of maximum of iterations for the EM algorithm.
N.random	The number of random sets of parameters for initial random starts.
show.random.ite	If "TRUE", the number of each iteration is shown on the R console.
eps	The criterion for the convergence of the alternating least-squares algorithm, which should be specified as a positive real value. If the difference between the values of penalized log likelihood functions of successive iteration is smaller than eps, then cbird makes a decision about the convergence of the algorithm.
mc.cores	If "parallel" package has been installed, "cbird" adopts a multithread process for multiple initial random starts. If "mc.cores"=1, "parallel" package is not needed, and a single core process is conducted.

Value

F	An estimated component score matrix for cluster centroids.
A	An estimated loading matrix.
mu	Estimated mean values in the subspace.
U	The cluster assignment matrix (N * N.clust).
g	The estimated mixture probability.
n.ite	The number of iteration needed for convergence.
loss	The value of log likelihood with L1 penalty.
bic	The value of BIC.
LL	The value of log likelihood.
cluster	Estimated clusters where subjects were assigned to
ptime	Time for calculation

Author(s)

Michio Yamamoto
[e-mail]

References

Yamamoto, M. and Hayashi, K. (2015). Clustering of multivariate binary data with dimension reduction via L1-regularized maximization. Pattern Recognition, 48, 3959-3968.

Examples

##Random Binary Data (unmeaningful example)
##100 subjects and 20 variables
##Consider three mixture components in the data
set.seed(1)
Y <- matrix(rbinom(100 * 20, 1, 0.5), 100, 20)
out <- cbird(Y, 2, 3)

est <- EstScore(Y, out$A, out$mu)

Example output

cbird documentation built on May 2, 2019, 2:42 a.m.