Parallelization experience with four canonical econometric models using ParMitISEM

Nalan Bastürk, S. Grassi, L. Hoogerheide, Herman K. van Dijk

Research output: Contribution to journalArticleAcademicpeer-review

116 Downloads (Pure)

Abstract

This paper presents the parallel computing implementation of the MitISEM algorithm, labeled Parallel MitISEM. The basic MitISEM algorithm provides an automatic and flexible method to approximate a non-elliptical target density using adaptive mixtures of Student-t densities, where only a kernel of the target density is required. The approximation can be used as a candidate density in Importance Sampling or Metropolis Hastings methods for Bayesian inference on model parameters and probabilities. We present and discuss four canonical econometric models using a Graphics Processing Unit and a multi-core Central Processing Unit version of the MitISEM algorithm. The results show that the parallelization of the MitISEM algorithm on Graphics Processing Units and multi-core Central Processing Units is straightforward and fast to program using MATLAB. Moreover the speed performance of the Graphics Processing Unit version is much higher than the Central Processing Unit one.
Original languageEnglish
Article number11
Number of pages20
JournalEconometrics
Volume4
Issue number1
DOIs
Publication statusPublished - Mar 2016

Keywords

  • Importance sampling
  • parallel computing
  • MitISEM
  • MCMC
  • MONTE-CARLO
  • REDUCED RANK
  • POSTERIOR
  • COMPUTATION
  • INTEGRATION
  • SIMULATION
  • DENSITIES
  • CHAIN

Cite this

Bastürk, Nalan ; Grassi, S. ; Hoogerheide, L. ; van Dijk, Herman K. / Parallelization experience with four canonical econometric models using ParMitISEM. In: Econometrics. 2016 ; Vol. 4, No. 1.
@article{be83a8fb07904b14a8ab144e40f6ea01,
title = "Parallelization experience with four canonical econometric models using ParMitISEM",
abstract = "This paper presents the parallel computing implementation of the MitISEM algorithm, labeled Parallel MitISEM. The basic MitISEM algorithm provides an automatic and flexible method to approximate a non-elliptical target density using adaptive mixtures of Student-t densities, where only a kernel of the target density is required. The approximation can be used as a candidate density in Importance Sampling or Metropolis Hastings methods for Bayesian inference on model parameters and probabilities. We present and discuss four canonical econometric models using a Graphics Processing Unit and a multi-core Central Processing Unit version of the MitISEM algorithm. The results show that the parallelization of the MitISEM algorithm on Graphics Processing Units and multi-core Central Processing Units is straightforward and fast to program using MATLAB. Moreover the speed performance of the Graphics Processing Unit version is much higher than the Central Processing Unit one.",
keywords = "Importance sampling, parallel computing, MitISEM, MCMC, MONTE-CARLO, REDUCED RANK, POSTERIOR, COMPUTATION, INTEGRATION, SIMULATION, DENSITIES, CHAIN",
author = "Nalan Bast{\"u}rk and S. Grassi and L. Hoogerheide and {van Dijk}, {Herman K.}",
note = "Data Sources: Yahoo Finance, Angrist-Krueger income-education data, FED.",
year = "2016",
month = "3",
doi = "10.3390/econometrics4010011",
language = "English",
volume = "4",
journal = "Econometrics",
issn = "2225-1146",
publisher = "MDPI",
number = "1",

}

Parallelization experience with four canonical econometric models using ParMitISEM. / Bastürk, Nalan; Grassi, S.; Hoogerheide, L.; van Dijk, Herman K.

In: Econometrics, Vol. 4, No. 1, 11, 03.2016.

Research output: Contribution to journalArticleAcademicpeer-review

TY - JOUR

T1 - Parallelization experience with four canonical econometric models using ParMitISEM

AU - Bastürk, Nalan

AU - Grassi, S.

AU - Hoogerheide, L.

AU - van Dijk, Herman K.

N1 - Data Sources: Yahoo Finance, Angrist-Krueger income-education data, FED.

PY - 2016/3

Y1 - 2016/3

N2 - This paper presents the parallel computing implementation of the MitISEM algorithm, labeled Parallel MitISEM. The basic MitISEM algorithm provides an automatic and flexible method to approximate a non-elliptical target density using adaptive mixtures of Student-t densities, where only a kernel of the target density is required. The approximation can be used as a candidate density in Importance Sampling or Metropolis Hastings methods for Bayesian inference on model parameters and probabilities. We present and discuss four canonical econometric models using a Graphics Processing Unit and a multi-core Central Processing Unit version of the MitISEM algorithm. The results show that the parallelization of the MitISEM algorithm on Graphics Processing Units and multi-core Central Processing Units is straightforward and fast to program using MATLAB. Moreover the speed performance of the Graphics Processing Unit version is much higher than the Central Processing Unit one.

AB - This paper presents the parallel computing implementation of the MitISEM algorithm, labeled Parallel MitISEM. The basic MitISEM algorithm provides an automatic and flexible method to approximate a non-elliptical target density using adaptive mixtures of Student-t densities, where only a kernel of the target density is required. The approximation can be used as a candidate density in Importance Sampling or Metropolis Hastings methods for Bayesian inference on model parameters and probabilities. We present and discuss four canonical econometric models using a Graphics Processing Unit and a multi-core Central Processing Unit version of the MitISEM algorithm. The results show that the parallelization of the MitISEM algorithm on Graphics Processing Units and multi-core Central Processing Units is straightforward and fast to program using MATLAB. Moreover the speed performance of the Graphics Processing Unit version is much higher than the Central Processing Unit one.

KW - Importance sampling

KW - parallel computing

KW - MitISEM

KW - MCMC

KW - MONTE-CARLO

KW - REDUCED RANK

KW - POSTERIOR

KW - COMPUTATION

KW - INTEGRATION

KW - SIMULATION

KW - DENSITIES

KW - CHAIN

U2 - 10.3390/econometrics4010011

DO - 10.3390/econometrics4010011

M3 - Article

VL - 4

JO - Econometrics

JF - Econometrics

SN - 2225-1146

IS - 1

M1 - 11

ER -