Перегляд за автором "Huliiev, N."

Зараз показано 1 - 3 з 3
  • Зображення мініатюри
    Публікація
    Comparison of dataset oversampling algorithms and their applicability to the categorization problem
    (ХНУРЕ, 2023) Teslenko, D.; Sorokina, A.; Khovrat, A.; Huliiev, N.; Kyriy, V.
    The subject of research in the article is the problem of classification in machine learning in the presence of imbalanced classes in datasets. The purpose of the work is to analyze existing solutions and algorithms for solving the problem of dataset imbalance of different types and different industries and to conduct an experimental comparison of algorithms. The article solves the following tasks: to analyze approaches to solving the problem – preprocessing methods, learning methods, hybrid methods and algorithmic approaches; to define and describe the oversampling algorithms most often used to balance datasets; to select classification algorithms that will serve as a tool for establishing the quality of balancing by checking the applicability of the datasets obtained after oversampling; to determine metrics for assessing the quality of classification for comparison; to conduct experiments according to the proposed methodology. For clarity, we considered datasets with varying degrees of imbalance (the number of instances of the minority class was equal to 15, 30, 45, and 60% of the number of samples of the majority class). The following methods are used: analytical and inductive methods for determining the necessary set of experiments and building hypotheses regarding their results, experimental and graphic methods for obtaining a visual comparative characteristic of the selected algorithms. The following results were obtained: with the help of quality metrics, an experiment was conducted for all algorithms on two different datasets – the Titanic passenger dataset and the dataset for detecting fraudulent transactions in bank accounts. The obtained results indicated the best applicability of SMOTE and SVM SMOTE algorithms, the worst performance of Borderline SMOTE and k-means SMOTE, and at the same time described the results of each algorithm and the potential of their usage. Conclusions: the application of the analytical and experimental ethod provided a comprehensive comparative description of the existing balancing algorithms. The superiority of oversampling algorithms over undersampling algorithms was proven. The selected algorithms were compared using different classification algorithms. The results were presented using graphs and tables, as well as demonstrated in general using heat maps. Conclusions that were made can be used when choosing the optimal balancing algorithm in the field of machine learning.
  • Зображення мініатюри
    Публікація
    Study of prediction and classification models in the problems of diabetes among patients with a stroke in different living conditions
    (ХНУРЕ, 2023) Huliiev, N.; Peretiaha, M.; Khovrat, A.; Teslenko, D.; Nazarov, A.
    The subject of the study in the article is the methods of predicting the development of diabetes. Diabetes mellitus is a non-communicable disease that has affected 425 million people, and by 2045 the number will only increase by 1.5 times. It has been proven to be an independent contributing factor to stroke development. When there is too much sugar in the blood, it negatively affects the arteries and blood vessels. People with this disease are more likely to develop atherosclerotic plaques and blood clots, which can lead to heart blockage and ischemic stroke. Having diabetes increases the risk and worsens the course of a stroke. According to the Framingham Study, the number of recurrent cases doubles. The aim of the study is to investigate methods of predicting and classifying the development of diabetes among people, in particular stroke patients, to prevent the development of other diseases. The complexity of the problem lies in the fact that there are as many undiagnosed cases as diagnosed ones, so about half of people suffer from the disease and the resulting complications due to improper or delayed diagnosis. Therefore, timely diagnosis of a disease that is difficult to detect is important in order to prevent the development of further complications. The article solves the problem of a multi-criteria task of choosing the best algorithm for predicting the occurrence of a disease. The following methods are used in this paper: multilayer perceptron, k-nearest neighbors method, decision tree, and logistic regression. Nowadays, machine learning has begun to apply to similar problems. In the 1950s and 1960s, there were attempts to combine the approaches to creating neural networks that existed at the time, which made it possible to calculate quantitative descriptions of human intelligence, and memorize, analyze, and process information, which resembled the work of the human brain. Medicine is one of the main areas of human activity where various classifier and neural network algorithms are gaining popularity yearly. They are trendy in disease diagnostics. Results: the initial conditions for choosing the best model are met by logistic regression. Conclusions: as a result of the study, the optimal model for predicting the development of the disease was selected.
  • Зображення мініатюри
    Публікація
    Theoretical and experimental study of the economic behavior irrational component of ukrainian society in specific markets
    (ХНУРЕ, 2023) Khovrat, A.; Teslenko, D.; Huliiev, N.; Kyriy, V.
    The subject matter of the article is the theoretical-methodical and applied principles of behavioural economics and their implementation in Ukrainian society. The goal of the work is to analyse the theory of irrationality in the economic context to find out what its character is in modern Ukrainian conditions, as well as to confirm the main paradoxes inherent in the individual’s decision-making behaviour. The following tasks were solved in the article: highlighting important aspects of the theory of irrationality for experimental analysis; determination of the methodology of experiments based on internationally recognized works; proposing a hypothesis regarding Ukrainian realities; conducting experiments according to the proposed methodology to test the proposed hypotheses and systematize the obtained results. The following methods are used: analytical and inductive methods for determining the set of behavioural experiments and building hypotheses regarding their results; experimental method and method of mathematical processing to check the presence of selected behavioural deviations in Ukraine. The following results were obtained: it was determined that in Ukrainian society there is a difference between the degree of individualistic attitudes in different age groups; determined change in the perception of information of different generations as a result of more significant digitization of the young population; a higher tendency of children to risk for certain conditions is determined; the similarity of the obtained results of the experiments with the global ones are established and the impact of technologies on the economic behaviour of individuals and the peculiarities caused by the historical context and expanded access to information, in general, are determined. Conclusions: the use of analytical and inductive methods in combination with an experimental approach confirmed the existence of some of the classic behavioural patterns for modern Ukrainian society, in particular: the Allais paradox, the effect of bounded rationality, the effect of joining the majority, the effect of ownership and asymmetric dominance. In addition, based on the obtained results, it was determined that for Ukraine there is a significant difference between the nature of irrationality in different age groups, however, the postulation of this statement beyond the selected target groups requires additional research, as well as consideration of the context of other market entities.