TRANSPORT DEVELOPMENT MODEL CONSTRUCTING ON THE BASIS MULTIVARIATE STATISTICAL METHODS
|Head:||Irina Yatskiv, Dr.Sc.Ing., Prof.|
Ph.: +371-7100594, Fax: +371-7100535, E-mail: firstname.lastname@example.org
|Members:||Jelena Jurshevich, M.Sc.|
Mihails Savrasovs, M.Sc.
Nadezda Kolmakova, M.Sc.
Topic of Research
The present research has been devoted to the further development and perfection of the methods of statistical analysis in transport sphere.
The main research activities within this frame are:
- Forecasting the transport development and transport flows
- Indicators of transport development constructing
Nowadays the tendencies of increasing in transportation flow make high demands for transport flow forecasting precision. It concerns any type of flow: cargo, passengers, air, rail, city traffic flow etc. The first direction of research is based on using the regressions models and time series analysis for forecasting the transport flow.
The second direction of research is connected with the following items: trying to construct the statistical models in the transport development area, the researchers often face the problem of considerable number of independent variables and small number of observations. The selection of only a part of them leads to the distortion of a the real situation. It is also linked with the regression models of transport development: investigating the problems of one type of transport, we cannot consider it apart from others, because the modern transport presents one system in the social and economic respects. So, the first problem - the large number of independent variables and the small number of observations in econometric models of transport development. The second problem - the significant correlations between the indices of transportation development. It troubles the construction of corrective statistical models.
For solving these problems the factor analysis methods can be used on the preliminary step. The factor analysis helps to form more compact set of data on the basis of the initial variables and other technologies.
Also this direction concerns with possibility of constructing the scalar index which can be used for estimating the transport development the regions and countries on the basis of various considered characteristics. The discriminant and cluster analyses are used for it.
This investigation gives the possibility and provides the methodology for using the set of multivariate statistical procedures for transport development study.
1. Yatskiv I. Solution of Dimension Decreasing Problem on the Transport Development Indicators Basement. Computer Modelling & New Technologies, 2002, Volume 6(2), pp. 21-34. - (in Russian)
2. Yatskiv I., Petuhovs E. Transport Development Statistical Model Constructing // Proceedings of the 7th International Conference "TransBaltica-2002". 14-15 June 2002, Riga Latvia. - Riga, Latvia, pp. 191-200.
3. Yatskiv I. Dimensionality reduction problem in the statistical analysis tasks of transport development data. // The 7-th Tartu Conference on Multivariate Statistics. Satellite meeting of the ISI 54th session in Berlin, 7-12 August 2003, Tartu, Estonia. - Tartu, 2003, p.58.
4. Vavilov V., Yatskiv I. Classification algorithm with covering. / Transport and Telecommmunication, 2002, Vol. 3 (4), pp. 92-95. - (in Russian).
5. Gusarova L., Yatskiv I. Methods of Identification of the Number of Clusters with Classification without Teaching. / Transport and Telecommmunication, 2003, Vol. 4 (1), pp. 23-28. - (in Russian).
6. Yatskiv I. Dimension Reduction Problem in the Statistical Analysis Tasks of Transport Development Data. Proceeding of XXIV International Seminar on Stability Problems for Stochastic Models. 13-17 September 2004, Jurmala, Latvia. pp. 345-352.
7. Yatskiv I. Gusarova L. The methods of checking the results of the iterative classification. \\The international Conference ?Reliability and Statistics in Transport and Communication- 2004", 14-15 October 2004, Riga, Latvia. Rīga, TSI, 2004. p.15.
8. Yatskiv I. Gusarova L. The different approaches to validation of cluster analysis results. The International Symposium on Stochastic Models in Reliability, Safety, Security and Logistics. 15-17 February 2005, Beer Sheva - Beer Sheva. p. 112.
9. D. Santalova, I.Jackiva. Factoru analīzes izmantošana transporta nozares attīstības tendenču noteikšanas ES valstīs / Mašinzinatņu un transports. Intelektuālās transporta sistēmas. RTU, Riga. 2005. p. 139-147
10. I.Yatskiv I. M.Havkunov. Air traffic flow forecasting models development for Riga flight information region. \\ Proceeding of International Conference "Reliability and Statistics in Transport and Communication - 2006", (RelStat'06). 25-28 October 2006, Riga, Latvia Vol.8, -1. Rīga, TSI, 2006. p.12-31. - (in Russian).