Decision tree method for identification and classification of information signals

• Дата: 2022
• Автори: Волошко, А. В., Джеря, Т. Е.
• Формат: Стаття
• Мова: Ukrainian
• Опубліковано: Інститут проблем реєстрації інформації НАН України 2022
• Теми: electrical load graph; information signal; multidimensional modeling; power consumption; decision tree; wavelet transform; wavelet basis
• Онлайн доступ: http://drsp.ipri.kiev.ua/article/view/275079
• Назва журналу: Data Recording, Storage & Processing

Recently, the issue of the development of intelligent active-adaptive electrical networks in the energy industry is often considered. Smart electrical networks have many different aspects. The uncertainty of information is one of them and is characterized by insufficiency, unreliability, ambiguity and uncertainty, and, in addition to physical factors, is associated with economic and temporal factors. During the functioning of electrical networks, it is necessary to qualitatively and correctly assess the degree of uncertainty in solving various problems of the energy sector. The article deals with the issue of correct classification of information signals using the decision tree method. The decision tree method allows you to understand and explain why a specific object belongs to one or another class. Packet wavelets are used to build a balanced wavelet transform tree. The algorithm of the method is indicated with a description and graphic drawings. The advantages of the chosen method and an example analysis are presented. The relevance of the question is determined by the fact that the process of assigning the electrical load schedule to a certain class is significantly accelerated with the help of the decision tree method. In the conclusion, an analysis of the work performed is carried out and a vector of future research is determined for optimization and more accurate results. The research, which was carried out using wavelet analysis, made it possible to model the multidimensional information flow with a complete and incomplete original data set.