In the article, a problem of prediction a user's psychological response to the presented image is considered. A complex algorithm that solves proposed problem with respect to fuzzy input data which operates in real time is proposed. A classification problem for each particular user based on previously generated content by the user using a set of algorithms including machine learning algorithm is considered. Due to a huge amount of variety in possible input data and it’s complexity algorithms that reduce fuzziness are considered. In order for a prediction system to understand and learn from data it has been provided, it has to be prepared in such a way that the algorithm could more easily find patterns and inferences. For that purpose, all incoming data passes two additional steps which also allows the system to neglect fuzziness. At the first step, the software has to definitely determine a presence of the desired object (face, in the article) which must be done fast enough to be made in a time of present-ing the picture to a user and it must be stable to data fuzziness. In order to solve that problem histogram of oriented graphs is considered. Because of its nature, the algorithm solves a problem in comparatively fast and robust way. At the second step, the face landmark estimation algorithm is considered. It allows the system to reduce the number of patterns that have to be learned in order to predict the behavior reducing the number of such patterns by narrowing down the two-dimensional transform of the object. The last step, prediction, is implemented with an artificial neural network that embraces incoming object and determines user’s resolution based on its experience. Such algorithm allows the system to keep learning throughout its life cycle which leads to constant im-provements in its results. The proposed algorithms complex was implemented. High speed, fuzziness resistance, independence of external conditions can be considered as its advantages. Slow learning cycle can be considered as its main disadvantage. Tests results provided in the article

Розглянуто проблему автоматизації обліку руху здобувачів вищої освіти Дніпровського національного університету імені Олеся Гончара. Розроблена автоматизована система для обліку та прогнозування контингенту студентів. Ключові слова: інформаційні технології; інформаційні системи; бази даних; контингент студентів; прогнозування контингенту студентів; облік студентів; мова С#. Рассмотрена проблема автоматизации учета движения соискателей высшего образования Днепровского национального университета имени Олеся Гончара. Разработана автоматизированная система для учета и прогнозирования контингента студентов. Ключевые слова: информационные технологии; информационные системы; контингент студентов; прогнозирование контингента студентов; учет студентов; язык С#.