Kurinienko O. Modeling the Probability of Business Bankruptcy Based on the Theory of Fuzzy Logic.

The dissertation work presents the elaboration of theory, methodical and practical aspects concerning modeling the probability of business bankruptcy on the basis of the theory of fuzzy logic. Scientific and methodical approaches have been improved regarding the following issues: analyzing financial reporting depending on available methods and information base at an enterprise; the choice of discriminative models for assessing the probability of bankruptcy from the point of view of their emphasis on obvious signs of a crisis in one-dimensional and multidimensional analysis as well as a simplified statistical procedure; tools of transforming factors of selected models according to New Accounting Regulations (Standards) with the purpose of extending the period of the object diagnostic. Three groups of models for assessing a business bankruptcy probability have been studied with their further detailed classification. Having synthesized their advantages and disadvantages, we have refined the complex scientific and methodical approach to the selection of discriminative models for assessing bankruptcy probability (their one-dimensional and multidimensional analysis and acceptable statistical procedures). Detailed consideration of discriminative analysis models was conducted with the aim of their further testing at Ukrainian enterprises in a particular industry. An approach to assessing the probability of bankruptcy has been developed, the main idea of which is ranking integral assessments of discriminative models using psychophysical "desirability scale" through preliminary valuation of key function indicators, their transformation into partial and generalized values. Evaluation results have been adjusted on the basis of this approach. The modeling of diagnostics and bankruptcy probability prediction has been automated. The developed software product has enabled to automate the use of financial reporting data of enterprises thus accelerating the analysis of creditworthiness of business partners and competitors. Approaches to optimizing parameters in modeling of any other processes in the context of actual data from the work of financially reliable business entities and potential outsiders have been theoretically grounded. Moreover, the proposed toolkit can be applied in other areas and fields.