Cluster Analysis Russian Federation Regions by Demand for Transport Services

Abstract

The article presents the application of cluster analysis, one of the most common machine learning methods, to the study of the level of regional transport development, segmentation of regions by the existing demand for transport services, and determination of the most important components of the transport system for certain regions. The purpose of the study is to divide the regions of the country into clusters that are relatively homogeneous in the main aspects of demand for transport services. Each cluster combines regions with similar economic, geographical and economic characteristics, which determines similarity in the most demanded modes of transport and transport infrastructure facilities. The research methodology is based on machine learning algorithms, the application of mathematical metrics to sets of statistical data. The research includes: selection of significant factors; analysis and normalization of statistical data; various clustering methods. In normalization, the data are converted to a single scale from 0 to 100 points with outliers excluded. As a result of cluster analysis, regions are distributed into four main clusters. Technical implementation of different variants of cluster analysis is possible in tabular editors and statistical packages. Based on the clustering results, each cluster is interpreted and common characteristics of the regions within them are identified. The prospects of the study include its annual updating based on updated statistical data. The results can be used to: analyze and develop the country’s transport system; identify priorities in the development of regional transport infrastructure; assess the significance and necessity of regional transport projects.

Affiliated Institutions

• Financial University RU

Related Publications