Requirements for intelligent data analysis systems and their classifications

DOI: 10.31673/2412-4338.2019.013136

Abstract

The article deals with the main existing diverse data analysis systems and their classification, which are used as a mass product for business applications or for unique research. Today, the application of not all systems is optimal, especially when it comes to the speed of data processing. The work analyzes the requirements relating to the systems of data mining, namely, support for work with data of large volumes, support for work with data that are heterogeneous in quality composition, providing work with noisy data, ensuring the availability of one mathematical algorithm for solving problems. the tasks that belong to various problem areas and to ensure the ease of work with a program of specialists without additional mathematical knowledge. Intelligent processing of data was formed at the junction of areas such as applied statistics, pattern recognition, database theory, artificial intelligence, and others, which explains a large number of methods and algorithms implemented in various existing systems of data mining, in addition some of them integrate several approaches at a time. On the basis of the analysis, we arrive at the conclusion that the system with the given problem is the most complete with the help of neural networks. Such a system can work with noisy data, data of a large volume. In addition, the system based on neural networks can be applicable to a sufficiently wide range of tasks and does not require the user to have special knowledge, as the network setup process is replaced by the learning phase. Also, the system contains a single mathematical device that does not require special knowledge of the user. Such systems have one significant minus - the results are often not easily interpreted.

Key words: intelligent data analysis, data mining system, neural network.

References (MLA)

1. Storchak K.P., Tushych A.M., Kozelkova K.S., Stepanov M.M. “Intelligent analysis of data using neural networks.” Communication 4 (2008): 17-19. Print.
2. Tushych A.M. “Analysis of the feasibility of using automated data mining system based on artificial neural networks.” VІІ All-Ukrainian Scientific and Practical Conference of Students, Postgraduates and Young Scientists on Automatic Control. Kherson. (April 10-12, 2019).:76-77.
3. Kulakov P.A. “Basic classes of neural networks in the problem of diagnostics of technological equipment.” Electronic scientific and practical magazine "Youth scientific journal" 10 (2016): 74-77.
4. Dyuk V.A. “Data Mining - Intelligent Analysis of Data.” Information technology: website. URL: http://www.inftech.webservis.ru/it/database/datamining/arhtml (date of request 03.12.2018)
5. Nazarov A.V. Loskutov A.I. Neural network algorithms for prediction and optimization of systems. Science and Technology, 2015. Print.
6. Lienkov S.V., Dzhulii B.M., Horbatiuk O.M., Bernaz H.M. “Conceptual scheme of the system of intellectual data processing.” Collection of scientific works of the Military Institute of Kyiv National Taras Shevchenko University 46 (2014): 181-190.
7. Xianjun N. “Research of Data Mining Based on Neural Networks.” World Academy of Science, Engineering and Technology 15 (2008): 381-384. Print.
8. Yang J., Parikh D., Batra D. “Joint unsupervised learning of deep representations and image clusters.” In CVPR (2016): 5147–5156. Print.
9. Yang B., Fu X., Sidiropoulos N. D., Hong M. “Towards k-means-friendly spaces: Simultaneous deep learning and clustering.” arXiv preprint arXiv:04794. (2016). Print.
10. Li G., Hongjun. L. “Data warehouse and data mining.” Microcomputer Applications 15(9). (1999): 17-20. Print.
11. Hand D., Mannila H., Smyth P. Principles of Data Mining. The Massachusetts Institute of Technology Press, 2001. Print.