TY - JOUR
T1 - SQUAREM
T2 - An R package for off-the-shelf acceleration of EM, MM and other EM-like monotone algorithms
AU - Du, Yu
AU - Varadhan, Ravi
N1 - Publisher Copyright:
© 2020, American Statistical Association. All rights reserved.
Copyright:
Copyright 2020 Elsevier B.V., All rights reserved.
PY - 2020
Y1 - 2020
N2 - We discuss the R package SQUAREM for accelerating iterative algorithms which exhibit slow, monotone convergence. These include the well-known expectation-maximization algorithm, majorize-minimize (MM), and other EM-like algorithms such as expectation conditional maximization, and generalized EM algorithms. We demonstrate the simplicity, generality, and power of SQUAREM through a wide array of applications of EM/MM problems, including binary Poisson mixture, factor analysis, interval censoring, genetics admixture, and logistic regression maximum likelihood estimation (an MM problem). We show that SQUAREM is easy to apply, and can accelerate any fixed-point, smooth, contraction mapping with linear convergence rate. The squared iterative scheme (SQUAREM) algorithm provides significant speed-up of EM-like algorithms. The margin of the advantage for SQUAREM is especially huge for high-dimensional problems or when the EM step is relatively time-consuming to evaluate. SQUAREM can be used off-the-shelf since there is no need for the user to tweak any control parameters to optimize performance. Given its remarkable ease of use, SQUAREM may be considered as a default accelerator for slowly converging EM-like algorithms. All the comparisons of CPU computing time in the paper are made on a quad-core 2.3 GHz Intel Core i7 Mac computer. R package SQUAREM is available from the Comprehensive R Archive Network (CRAN) at https://CRAN.R-project.org/package=SQUAREM/.
AB - We discuss the R package SQUAREM for accelerating iterative algorithms which exhibit slow, monotone convergence. These include the well-known expectation-maximization algorithm, majorize-minimize (MM), and other EM-like algorithms such as expectation conditional maximization, and generalized EM algorithms. We demonstrate the simplicity, generality, and power of SQUAREM through a wide array of applications of EM/MM problems, including binary Poisson mixture, factor analysis, interval censoring, genetics admixture, and logistic regression maximum likelihood estimation (an MM problem). We show that SQUAREM is easy to apply, and can accelerate any fixed-point, smooth, contraction mapping with linear convergence rate. The squared iterative scheme (SQUAREM) algorithm provides significant speed-up of EM-like algorithms. The margin of the advantage for SQUAREM is especially huge for high-dimensional problems or when the EM step is relatively time-consuming to evaluate. SQUAREM can be used off-the-shelf since there is no need for the user to tweak any control parameters to optimize performance. Given its remarkable ease of use, SQUAREM may be considered as a default accelerator for slowly converging EM-like algorithms. All the comparisons of CPU computing time in the paper are made on a quad-core 2.3 GHz Intel Core i7 Mac computer. R package SQUAREM is available from the Comprehensive R Archive Network (CRAN) at https://CRAN.R-project.org/package=SQUAREM/.
KW - Convergence acceleration
KW - EM algorithm
KW - Extrapolation methods
KW - Fixed-point iteration
KW - High dimensional models
KW - Monotone convergence
KW - Optimization
UR - http://www.scopus.com/inward/record.url?scp=85079840287&partnerID=8YFLogxK
UR - http://www.scopus.com/inward/citedby.url?scp=85079840287&partnerID=8YFLogxK
U2 - 10.18637/jss.v092.i07
DO - 10.18637/jss.v092.i07
M3 - Article
AN - SCOPUS:85079840287
VL - 92
SP - 1
EP - 41
JO - Journal of Statistical Software
JF - Journal of Statistical Software
SN - 1548-7660
ER -