Forecasting the Risk of Bankruptcy in Poorly Formalized Processes

There is a class of complex systems characterized by dynamism, multi-link structural elements, multi-stage, multi-linked chain of processes. Moreover, each of these processes occurs under conditions of stochastic and non-stochastic uncertainty in the initial information, internal and external environment, which predetermine the uncertainty of the nature of the development of the situation. Decision-making problems in such systems are divided into two types: 1) decision-making problems under risk conditions, when uncertainty conditions are only probabilistic, stochastic in nature; 2) decision-making problems under conditions of uncertainty, when the accompanying conditions are of a non-stochastic nature, and also when the necessary reliable statistical data is unknown. In tasks of the second type, risks are manifested to a greater extent than in the first. At the same time, risk should be considered – as an object, event, phenomenon – as a formal mathematical category in accordance with its following information interpretation: risk is information uncertainty, fuzziness of the “object – subject – environment” system and its individual elements. The measure of this uncertainty determines the measure of danger, possible damage, loss from the implementation of some decision or event. The existence of risk is associated with the inability to predict the future with 100 % accuracy. Based on this, the main property of risk should be singled out: risk occurs only in relation to the future and is inextricably linked with forecasting, and therefore with decision-making in general (the word “risk” literally means “making a decision”, the result of which is unknown). Following the above, it is also worth noting that the categories “risk” and “uncertainty” are closely related and are often used as synonyms. In conditions when the initial factors are given in the form of fuzzy characteristics, other approaches based on the intelligent technologies of Soft Computing are widely used for forecasting. When evaluating alternative decision-making options for risk assessment under uncertainty, the problem of developing fuzzy models based on fuzzy inference rules arises. But there is no universal method for constructing fuzzy evaluation models. The advantage of fuzzy logic lies in the possibility of using expert knowledge about a given object in the form of if “inputs”, then “outputs”. In the paper a bankruptcy risk model is developed in poorly formalized processes for the purpose of forecasting.