Use of Kohonen’s Self-Organizing Maps and the Monte Carlo Method for Studying Adequacy of Factor Models of Intelligence

767

Abstract

The article presents new methods of identification and research of factors which determine the development of intellectual abilities. It is emphasized that its peculiar quality is construction of predetermined system of linear and non-linear algebraic equations in relation to the model’s free parameters with consequent search of a pseudosolution. Identification of free parameters of the factor model is done by means of least square method. A method of evaluation of statistical importance of factor model components is considered as well as a new method of evaluation of level of adequacy of random factor models which is based on the Monte Carlo method and capabilities of Kohonen’s self-organizing maps. Attention is drawn to the fact that this method allows to avoid strict limitations on probability distributions of observations results, which are peculiar to the traditional procedure of identification of model’s free parameters. The article presents advantages of this approach over the traditional method as well as a number of factor models presented by way diagrams including their comparative analysis.

General Information

Keywords: factor analysis, adequacy of factor models, Kohonen’s self-organizing feature maps

Journal rubric: Interdisciplinary Researches

Article type: scientific article

For citation: Panfilova A.S. Use of Kohonen’s Self-Organizing Maps and the Monte Carlo Method for Studying Adequacy of Factor Models of Intelligence. Psikhologicheskaya nauka i obrazovanie = Psychological Science and Education, 2011. Vol. 16, no. 5, pp. 88–99. (In Russ., аbstr. in Engl.)

References

  1. Galushkin A. I. Nejronnye seti. Osnovy teorii. M., 2010.
  2. Ermakov S. M. Metod Monte-Karlo v vychislitel'noj matematike. Vvodnyj kurs. SPb., 2009.
  3. Kuravskij L. S., Baranov S. N., Malyh S. B. Nejronnye seti v zadachah prognozirovanija, diagnostiki i analiza dannyh. M., 2003.
  4. Kuravskij L. S., Marmaljuk P. A., Panfilova A. S., Ushakov D. V. Issledovanie faktornyh vlijanij na razvitie psihologicheskih harakteristik s primeneniem novogo podhoda k ocenke adekvatnosti modelej nabljudenijam // Informacionnye tehnologii. 2011. № 11. (V pechati).
  5. Kuravskij L. S., Marmaljuk P. A., Baranov S. N., Abramochkina V. I., Petrova E. A. Faktornyj analiz rezul'tatov vejvlet-preobrazovanij longitjudnyh dannyh kak novyj metod issledovanija dinamicheskih harakteristik slozhnyh sistem // Nejrokomp'jutery: razrabotka i primenenie. 2009. № 9.
  6. Kuravskij L. S., Marmaljuk P. A., Abramochkina V. I., Petrova E. A. Primenenie faktornogo analiza rezul'tatov vejvlet-preobrazovanij dlja issledovanija dinamiki psihologicheskih harakteristik // Eksperimental'naja psihologija. 2009. № 1.
  7. Marmaljuk P. A. Ocenka stepeni adekvatnosti faktornyh modelej c pomosh'ju samoorganizujushihsja kart priznakov Kohonena // Nejrokomp'jutery: razrabotka i primenenie. 2010. № 10.
  8. Ushakov D. V. Intellekt: strukturno-dinamicheskaja teorija. M., 2003.
  9. Kuravsky L. S., Marmalyuk P. A., Panfilova A. S. Estimation of goodness-of-fit measures for identification of unrestricted factor models employing arbitrarily distributed observed data // In: Proc. 8th International Conference on Condition Monitoring & Machinery Failure Prevention Technologies. Cardiff, UK, June 2011.
  10. Neale M. C., Cardon L. R. Methodology for genetic studies of twins and families. Dordrecht, the Netherlands, Kluwer Academic Publishers, 1992.

Information About the Authors

Anastasya S. Panfilova, PhD in Engineering, Researcher, Institute of Psychology, Russian Academy of Sciences, Moscow, Russia, ORCID: https://orcid.org/0000-0003-1892-5901, e-mail: panfilova87@gmail.com

Metrics

Views

Total: 1625
Previous month: 8
Current month: 0

Downloads

Total: 767
Previous month: 2
Current month: 0