Intellectual analysis of data using neural networks
DOI №______
Abstract
The article considers artificial neural networks as a means of intellectual data analysis. The urgency of using the proposed data analysis means lies in the active growth of data as a result of automation of various technological processes. The main methods for performing data analysis are analyzed in the paper, and their advantages and disadvantages are identified. As a result, it was concluded that the use of artificial neural networks is expedient. The main advantage of artificial neural networks usage is the ability to solve various informational tasks. The use of such networks for data analysis is advisable because they have properties of function approximation, learning ability, improvement of their own structure, low probability of error with correct initial setup of network parameters, analysis capabilities even in the presence of incomplete and noisy data. At the same time, it is possible to simulate the situation simply by submitting the data to the input of the network and evaluating the result that it issues. In the work the model of the system is constructed and the process of the intellectual data analysis is based on artificial neural networks and stages of its implementation. The proposed algorithm for solving the problem adequately describes the process of performing an analysis that goes through three stages: data preparation (does not require complex manipulation associated with digitization, preprocessing data is an extension of the data purification process), data analysis (the neural network can only work with numerical data, which implies that it is necessary to convert symbol data into numerical ones by creating correspondence tables between symbolic data and numeric or accepting hash-functions for creation unique numeric data), Expression (transformation data after pretreatment in the form which can be adopted subject to the data analysis algorithm) and interpretation of results (in user-friendly form).
Keywords: data mining; artificial neural network; algorithm.
References
1. Кулаков П. А. Основные классы нейронных сетей в задаче диагностики технологического оборудования // Молодежный научный вестник: электронный науч.-практ. журн. 2016. № 10. С. 74–77.
2. Манжула В. Г., Федяшов Д. С. Нейронные сети Кохонена и нечеткие нейронные сети в интеллектуальном анализе данных // Fundamental research. 2014. №4. С. 108–114.
3. Бирюков Е. В., Корнев М. С. Практическая реализация нечеткой нейронной сети при краткосрочном прогнозировании электрической нагрузки // Портал магистров ДонНТУ. URL:
http://masters.donntu.org/2006/kita/chuykov/links/index.htm (дата звернення 02.10.2018).
4. Дюк В. А. Data Mining — интеллектуальный анализ данных // Информационные технологии. URL: http://www.inftech.webservis.ru/it/database/datamining/ar2.html (дата звернення 01.11.2018)
5. Назаров А. В., Лоскутов А. И. Нейросетевые алгоритмы прогнозирования и оптимизации систем. СПб.: Наука и Техника, 2015. 384 c.
