Algorithms for decision support in risk management in the development of information-sensitive socially oriented systems

The increase in the amount of data and the overall change in social and market processes are driving the transformation of basic management principles. This is especially true for risk management. Under non-deterministic conditions, when developing solutions for socially oriented systems, forecasting possible problems requires the use of modern data mining tools. Existing concepts cannot fully guarantee high efficiency in the face of social shifts exacerbated by a falsified information environment, for example, during discrediting campaigns against ideas proposed by businesses. At the same time, there is a problem with the speed of simple solutions and the high cost and security of using more complex cloud technologies. The current work is focused on considering modifications to simple decision support models and building a data processing algorithm to improve the accuracy and reliability of project forecasting. The proposed sequence of steps allows us to take into account the essence of instability and its impact on society, while taking into account the peculiarities of the information environment. Experimental comparison of various classical models with the proposed solution allows us to state the higher efficiency of the created algorithm, both in terms of accuracy and speed due to parallelization. This, in turn, opens the way to solving the problem of risk prediction for information-sensitive socially oriented systems