| Authors |
Mikhail Yu. Mikheev, Doctor of technical sciences, professor, head of sub-department of informational technologies and systems, Penza State Technological University (1а /11 Baydukov passage /Gagarin street, Penza, Russia), Е-mail: mix1959@gmail.com
Oleg V. Prokof'ev, Candidate of technical sciences, associate professor of sub-department of mathematics and computer science, Penza State Technological University (1а /11 Baydukov passage /Gagarin street, Penza, Russia), E-mail: prokof_ow@mail.ru
Irina Yu. Semochkina, Candidate of technical sciences, associate professor, associate professor of sub-department of informational technologies and systems, Penza State Technological University (1а /11 Baydukov passage /Gagarin street, Penza, Russia), E-mail: ius1961@gmail.com
|
| Abstract |
Background. In the process of evolution of Decision Support System (DSS) computer systems, the "classic" decision trees were supplemented with new ways of displaying data – tree maps. Using a hierarchical structure, tree maps provide meaningfully organized displays of large amounts of information, which gives advantages in developing solutions for big data.
Materials and methods. Gathering and summarizing the experience of using ordinary trees and tree maps, the authors set a goal to develop recommendations on the use of data visualization tools in applied areas, which improves the quality of decision support. If the usual approach has found effective application in machine learning, then tree maps are oriented to manual use by decision makers. In contrast to the first approach, which relies on objective numerical criteria for assessing quality, the effect of using tree maps is subjective and less obvious. For the second approach, user survey data were applied during the solution of professional tasks related to the analysis of web pages "manually" in the process of a full factorial experiment.
Results and conclusions. The average time for solving problems and expert estimates allowed us to draw conclusions about the advantages of a tree-like presentation of data and the potential for improving the quality of support for decisions made. In conclusion, the article substantiates the conclusion about the need for the integrated use of data visualization tools and a number of other technologies in the second generation DSS 2 products.
|
| Key words |
decision trees, treemaps, quality of visualization, quality of solution, experimental data
|
