Search academic texts

Information × Registration Number 0215U000188, 0114U007140 , R & D reports Title Electric load forecasting method using evolutionary neuro-fuzzy network of hybrid neuron-like units popup.stage_title Head Popov Sergiy Vitaliyovych, Registration Date 29-01-2015 Organization Kharkiv National University of Radioelectronics popup.description2 The project concerns development of a new method of electric load forecasting (first of all, short-term) under uncertainty using evolutionary neuro-fuzzy network of hybrid neuron-like units. Improvement of forecasting accuracy is achieved due to considering the influence of factors defined on different measurement scales (quantitative, ordinal, nominal, qualitative) and the incorporation of a priori information about the form of influence of these factors on electric load and the specific relationships between the processes of power consumption. New advanced learning methods for the neuro-fuzzy network of hybrid neuron-like units that are based on an evolutionary approach will not only optimize the network's parameters, but also its architecture. This improves the electric load forecasting accuracy, makes forecasts more interpretable and more resistant to errors in the input information. The forecasting models are able to adapt to the changes in the properties of the power consumption processes that may arise due to, e.g. changing economic environment in the market.5481 Product Description popup.authors Попов Сергій Віталійович popup.nrat_date 2020-04-02 Close