Journal "Software Engineering"
a journal on theoretical and applied science and technology
ISSN 2220-3397

Issue N4 2017 year

DOI: 10.17587/prin.8.170-176
The Visualization Methods for Cluster Analysis Results of Mechanical Engineering Components based on Neural Network
V. A. Kharakhinov, tes4obse@mail.ru, S. S. Sosinskaya, sosinskaya@mail.ru, Irkutsk National Research Technical University, Irkutsk, 664074, Russian Federation
Corresponding author: Sosinskaya Sophia S., Professor, Irkutsk National Research Technical University, Irkutsk, 664074, Russian Federation, E-mail: sosinskaya@mail.ru
Received on November 30, 2016
Accepted on January 11, 2017

This paper considers the research results of clustering analysis based on Kohonen networks for the mechanical engineering components. Some ways to visualize these results were discussed in the paper. The first way is based on Andrews curves. It provides only an approximate graphical representation for this data. The next way uses factor analysis to reduce dimension of initial data and to visualize these data in lower dimension. The cluster analysis was applied to the factors values for time reducing of network training. A comparison of the results of cluster analysis of mechanical engineering components on the basis of self-organizing networks with outputs obtained using other clustering methods is given.

Keywords: cluster analysis, Kohonen networks, Andrews curves, factor analysis, mechanical engineering components
pp. 170–176
For citation:
Kharakhinov V. A., Sosinskaya S. S. The Visualization Methods for Cluster Analysis Results of Mechanical Engineering Components based on Neural Network, Programmnaya Ingeneria, 2017, vol. 8, no. 4, pp. 170—176.