MODELING SOCIOLOGICAL DATA WITH THE MARKOV CHAIN MONTE CARLO APPROACH

Journal Title: Jaunųjų mokslininkų darbai - Year 2014, Vol 42, Issue 2

Abstract

This work represents the Markov chain Monte Carlo method application for estimation of the multidimensional rare events frequencies and α- stable distribution. Multivariate MCMC algorithm, with a statistical termination rule and sample size regulation, and efficiency by modeling sociological data are researched in this paper. In this work the Monte Carlo sample size is regulated using statistical criteria when two Markov chains become equal by their Monte Carlo sample distribution. The 2003 Lithuanian population suicide and murders (homicide) data and five telecommunications companies’ stock data from 2012-01-20 to 2012-04-01 are used for modeling in the paper. The comparison of this method with a standard fixed-sample MCMC algorithm provided results that allows us to obtain a Poisson-Gaussian model and the α- stable law estimates with the required accuracy by less quantity of chains, and in our case allowed us to double to shorten the calculation time. The method for selecting the initial data for the Poisson-Gaussian model is proposed. Keywords: Markov chain Monte Carlo method, Bayesian approach, Poisson-Gaussian model, stable distribution, statistical modeling.

Authors and Affiliations

Ingrida Vaičiulytė, Leonidas Sakalauskas

Keywords

Related Articles

University Governance Models: the Case of Lapland University

It also presents the results of “International Comparative Research on University Governance Models” carried out by the paper authors emphasizing the context of Finland’s higher education and university governance and th...

MATHEMATICAL ESTIMATION OF THE AGE OF MOUNDS IN LITHUANIA

The aim of this paper is to develop a mathematical model for estimation of the age of mounds in Lithuania. On the grounds of archaeological literature it was found that mounds are of six periods. This study considers onl...

D klasės sportinių šokių šokėjų motyvacijos šokti ypatumai

Nustatyta, kad šokėjų (merginų ir vaikinų) vidinė motyvacija išreikšta labiau nei išorinė motyvacija. Iš vidinę motyvaciją atskleidžiančių motyvų labiausiai išreikšta vidinių motyvų grupė, susijusi su šokėjų siekiu "suži...

Download PDF file
  • EP ID EP90641
  • DOI -
  • Views 136
  • Downloads 0

How To Cite

Ingrida Vaičiulytė, Leonidas Sakalauskas (2014). MODELING SOCIOLOGICAL DATA WITH THE MARKOV CHAIN MONTE CARLO APPROACH. Jaunųjų mokslininkų darbai, 42(2), 181-187. https://www.europub.co.uk/articles/-A-90641