A dinamic method for risks evaluation in the financial management system

• Дата: 2019
• Автор: Kuznietsova, N. V.
• Формат: Стаття
• Мова: Ukrainian
• Опубліковано: Інститут проблем реєстрації інформації НАН України 2019
• Теми: динамічний метод; фінансові ризики; моделі виживання; модель пропорційних ризиків Кокса; оцінка Каплан-Майєра; dynamic method; financial risks; survival models; Cox proportional hazards model; Kaplan-Meier estimation
• Онлайн доступ: http://drsp.ipri.kiev.ua/article/view/183724
• Назва журналу: Data Recording, Storage & Processing

The main definitions and formalization of various types of survival models such as empirical survival function, generalized linear model, Cox proportional hazards model and its modifications and nonparametric models are presented for risk assessment. A dynamic method of risk assessment that allows to assess the risk’s degree and level and also to predict the transition from critical to catastrophic risk by using of parametric, semi-parametric and nonparametric models based on survival functions has been proposed. The method allows to apply data stratification and separately to simulate different survival functions for different data categories. The method involves the identification of statistically significant characteristics, the development of different survival models and selection of the best model by the set of criteria, testing the hypothesis about the same distribution of risk functions and at finally defined time, the probability of risk occurrence or probable losses are determined accordingly. Also it includes two developed by author algorithms that allow prediction of such a time based on the established permissible (critical) probability of risk occurrence or restriction of possible economic losses, in particular, the moment of risk transition from permissible to critical or catastrophic (in the defined amount of critical or catastrophic losses). The algorithm for calculating the moment of transition to the higher risk probability (risk degree) could be presented in two possible variations. The first possibility is defining time through the calculation of the derivative of the risk function, which was given explicitly. The second option is to define through the «probability reserve» (as the speed of probability change). The algorithm for determining the moment of occurrence of risk’s critical (catastrophic) level by losses is performed as an iterative step-by-step procedure of transition point search for losses from critical to catastrophic. Tabl.: 1. Fig.: 1. Refs: 12 titles.