matbaseERM: matrix-based Expectation-Regularization-Maximization(ERM)...

In ygu427/HDSSM: High-dimensional linear SSM for dynamic GRN

Description

A high-dimensional linear State Space Model(SSM) with a new Expectation-Regularization-Maximization(ERM) algorithm to construct the dynamic Gene Regularization Network(GRN). The new ERM algorithm employs the idea of the adaptive LASSO-based variable selection method to preserve the sparsity property of GRN.

Usage

matbaseERM(y, initA, initC, initQ, initR, initx, initV, max_iter = 100, diagQ = 0, diagR = 0, ARmode = 0, s.prop = 0.1^6, ...)

Arguments

y	y[,t] the observation vector at time t
initA	the initial system matrix
initC	the initial observation matrix
initQ	the initial variance for normally distributed system noise
initR	the initial variance for normally distributed measurement noise
initx	mean value vector for initial state x0
initV	covariance matrix for initial state x0
max_iter	specifies the maximum number of EM iterations (default 100)
diagQ	boolean value. 1 specifies that the Q matrix should be diagonal. Default value is 0,indicating fixed at true value.
diagR	boolean value. 1 specifies that the R matrix should be diagonal. Default value is 0,indicating fixed at true value.
ARmode	boolean value. 1 specifies that C=I, R=0. i.e., a Gauss-Markov process. (Default 0).
s.prop
...	more optional arguments

Value

estA	the estimated high-dimensional sparse system matrix A
estC	the estimated observation matrix
estQ	the estimated variance for normally distributed system noise
estR	the estimated variance for normally distributed measurement noise
estX	the mean value for state vector
estV	the covariance matrix for state vector
LL	the log likelihood vector
xcurve	smoothed real-valued hidden state variable vector
bic	BIC
num_iter	the number of iteration has been processed

Author(s)

Yu Gu

Maintainer: Yu Gu <[e-mail]>

References

Chen, J., & Chen, Z. (2008). Extended Bayesian information criteria for model selection with large model spaces. Biometrika, 95(3), 759-771.

Green, P. J. (1990). On use of the EM for penalized likelihood estimation. Journal of the Royal Statistical Society. Series B (Methodological), 443-452.

Harrison, J., & West, M. (1999). Bayesian Forecasting & Dynamic Models. Springer.

Zou, H. (2006). The adaptive lasso and its oracle properties. Journal of the American statistical association, 101(476), 1418-1429.

ygu427/HDSSM documentation built on May 4, 2019, 2:33 p.m.