Holub M. Formation processes of a numerical indications for ukrainian-language text’s classification in the intellectual monitoring technology

Dissertation is devoted to solve the scientific problem of increasing the adequacy of the ukrainian-language texts classification. It’s process combine developing methods and means of effective formation an array of their numerical indications for constructing model-classifiers in multi-level intellectual monitoring technologies. Information technology is implemented in the monitoring intellectual system (MIS) form. The result of the of the method's implementation of forming a dictionary of features, developing a new method for constructing observation’s points of the Ministry of Internal Affairs and coordinating them with the synthesizer of model-classifiers, the base of model knowledge and other elements of the MIS. It became possible to achieve systemic effect (immergence). Immergence is obtained in the form of ability to classify ukrainian-language text messages and increase the number of correctly classified texts compared to analogues. The share reached in 98-100%. At the same time, the size of a text message, categorized correctly, has been reduced to 500 characters. The results obtained during the study were confirmed experimentally in determining the place of residence of the authors of printed text messages, between the solving tasks of attribution and text profiling. The information technology of multi-level intellectual monitoring has acquired an ability to solve the problems of intellectual analysis of Ukrainian-language texts, in particular their classification.