Search academic texts

Information × Registration Number 0225U002657, (0122U000671) , R & D reports Title Development of models and methods for solving predictive problems based on large amounts of poorly structured information in conditions of uncertainty popup.stage_title Розробка моделей та методів розв’язання задач передбачення на основі великих обсягівслабкоструктурованої інформації в умовах невизначеності Head Pankratova Nataliia D., д.т.н. Registration Date 26-03-2025 Organization Educational and Scientific Complex "Institute for Applied System Analysis" of National Technical University of Ukraine "Igor Sikorsky Kyiv Polytechnic Institute" popup.description1 The aim of the work is to develop theoretical principles for predicting and creating a human-machine management decision-making system, providing recommendations based on morphological, hierarchical and hybrid neural network models and methods for large amounts of heterogeneous, including poorly structured information under uncertainty. popup.description2 The aim of the work is to develop theoretical foundations for solving foresighting problems and creating a human-machine management decision-making system, providing recommendations based on morphological, hierarchical and hybrid neural network models and methods when receiving large amounts of heterogeneous, including weakly structured information under conditions of uncertainty. A modified method of hierarchical and network models has been developed, which differs from others in calculating weights based on consistent and inconsistent multiplicative fuzzy and interval matrices of pairwise comparisons and fuzzy programming of preferences and leads to more reliable weights of elements of the decision-making model compared to other known methods, allows you to find the most inconsistent elements of the problem. Modified artificial intelligence models for decision-making and providing recommendations have been proposed and investigated, in particular, new models of recommender systems based on implicit preferences and neural network convolutional models for recognition tasks. The prediction methodology has been improved to build and systematically evaluate scenario alternatives when receiving large amounts of weakly structured data under conditions of uncertainty, which includes models, methods and approaches to structuring large amounts of weakly structured information with the removal of assessments, judgments and expectations described in natural language, decision-making support tools based on a clear and fuzzy apparatus of modified morphological analysis methods, modern techniques and models of text analytics for supporting scenario building tasks, modified methods based on hierarchical and network decision-making support models. A systematic approach to solving scenario analysis tasks has been developed with further modeling of the interaction process of qualitative analysis, neural network, fractal analysis and text analytics methods. Product Description popup.authors Androsov Dmytro V. Malyshevskyi Oleksii H. Nedashkivska Nadiia I. Pankratova Nataliia D. Savastianov Volodymyr V. Savchenko Illia O. popup.nrat_date 2025-03-26 Close