The predictive neural network model for business processes management 

Mikhail I. Krevskiy, Senior analyst, department of monitoring and analysis of public services, State government institution “Informatsionniy Gorod” (building 1, 21 Perviy Krasnogvardeyskiy passage, Moscow, Russia), E-mail: westhemer1@gmail.com
Aleksandr S. Bozhday, Doctor of engineering sciences, professor, professor of the sub-department of computer-aided design system, Penza State University (40 Krasnaya street, Penza, Russia), E-mail: bozhday@yandex.ru 

Background. Predictive business process analytics is an important tool in the management of organizational systems. The study discusses the approaches to improve deep learning methods for business processes prediction. The purpose of the work is to improve the efficiency of analysis and predictive identification of problematic business processes by improving existing deep learning methods. The implementation of the proposed methods will effectively predict the problems and the behavior of both single processes and entire organizational systems. Materials and methods. Process Mining methods, deep learning and methods of organizational system management are used. Results. In the course of the work, a review of existing process predicting methods was made. An improvement of existing solutions is proposed by integrating the properties of predictive neural networks and networks for machine translation within a single model. In particular, encoder and decoder were used, as well as the attention mechanism. Conclusions. The work experimentally confirmed the effectiveness of the proposed modification of the neural network architecture. The changes made it possible to improve the quality of predicting the business processes activities by 1.5–2 times. 

business process, machine learning, neural network architectures, methods for predicting the behavior of processes, long short-term memory, the mechanism of attention, predictive model 

Krevskiy M.I., Bozhday A.S. The predictive neural network model for business processes management. Izvestiya vysshikh uchebnykh zavedeniy. Povolzhskiy region. Tekhnicheskie nauki = University proceedings. Volga region. Engineering sciences. 2023;(3):83–93. (In Russ.). doi: 10.21685/2072-3059-2023-3-6