Вестник Научно-исследовательского института железнодорожного транспорта (Oct 2017)

Investigation of the impact of service parameters on the degree of passenger satisfaction based on the application of the apparatus of neural networks

  • S. S. Pastukhov,
  • K. V. Stel’Mashenko

DOI
https://doi.org/10.21780/2223-9731-2017-76-5-273-280
Journal volume & issue
Vol. 76, no. 5
pp. 273 – 280

Abstract

Read online

Based on the analysis of the advantages and disadvan-luation models has been identified and justified. These models could tages of various methodological approaches to the search for ways be used to guide management to implement several areas of service to increase the quality of service existing in the marketing practice of improvement at the same time, and also to make efficient decisions in the passenger complex, the need for further development of such the context of possible synergies between the multiple aspects of the robust nonparametric mechanisms for constructing predictive va-service being investigated in their impact on customer satisfaction. Methodical and practical approaches are proposed to identify and compare the effect of service parameters on the level of passenger satisfaction on the basis of modeling using the apparatus of neural networks for conditions where the application of least squares correlation and regression analysis and mechanisms of order logistic regression are prognostically inefficient or excessively time consuming for analytical work. The approbation results of the apparatus of neural networks on sets of marketing data concerning various aspects of passenger service at stations and in trains are presented, as well as an estimation of the efficiency of using neural networks in comparison with the algorithms of ordinal logistic and linear regression. As a result, in order to obtain the maximum of objective information necessary for developing solutions in the field of quality management of transport services of the passenger complex, it is proposed to use a new modification of the universal complex mechanism for studying the influence of various service parameters on the level of customer satisfaction, including a set of regression algorithms, the apparatus of neural networks and nonparametric statistics toolkit.

Keywords