Teaching a neural network modeling socio-economic development of the region

The article is devoted to the formation of an array of data for the construction of an artificial neural network, designed to search for relationships between social and economic parameters of the development of regions of the Russian Federation. The relevance of research in this area is confirmed both by a large number of studies in the field of regional comparativistics and by the limited methods used in this kind of research, often limited to descriptive methods and basic techniques of parametric statistics. Under these conditions, the expansion of the mathematical apparatus and the more active introduction of information technologies (including in the area of Big Data analysis and the construction of predictive models based on artificial neural networks) can be viable. At the same time, however, it should be noted that the resources of an individual research team may be (and most likely will be) insufficient to create their own software solution for the implementation of machine learning algorithms from scratch. The use of third-party cloud-based software platforms (primarily IBM and Google infrastructures) allows to bypass the problem of the research team’s lack of expensive material and technical base, however they impose a number of limitations dictated by the requirements of the existing machine learning algorithms and the specific architecture provided platforms This puts the research team in front of the need to prepare the accumulated data set for processing: reducing the dimension, checking the data for compliance with the platform requirements and eliminating potential problem areas: “data leaks”, “learning distortions” and others. The paper was reported to the section “Sociology of Digital Society: Structures, Processes, Governance” of the International Conference Session “Public Administration and Development of Russia: National Goals and Institutions”.

1. Aivazyan S.A., Bukhshtaber V.M., Enyukov I.S., Meshalkin L.D. (1989), Applied statistics: classification and dimension reduction [Prikladnaya statistika. Klassifikatsiya i snizhenie razmernosti], in Aivazyan S.A. (ed.), Finansy i statistika, Moscow, Russia. [in Russian].

2. Vasenkov D.V. (2007), “Methods of teaching artificial neural networks” [“Metody obucheniya iskusstvennykh neironnykh setei”], Computer tools in education, no. 1, pp. 20–29.

3. Tolstova Yu.N. (2015), “Sociology and Computer Technologies” [“Sotsiologiya i komp’yuternye tekhnologii”], Sotsiologicheskie issledovaniya, no. 8, pp. 3–13.

4. James G. (2003), “Variance and Bias for General Loss Functions”, Machine Learning. Available at: http://www-bcf.usc.edu/~gareth/research/bv.pdf (accessed 10.07.2019).

5. Mohri M., Rostamizadeh A., Talwalkar A. (2012), Foundations of Machine Learning, The MIT Press, Cambridge, USA.