Mathematical Modeling of the Evolution of the Rehabilitation Process for Patients with Oncological Diseases

This study is focused on addressing the issue of planning the progression of the cancer rehabilitation process. The rehabilitation of cancer patients is an intricate, multi-stage, and multifactorial procedure that necessitates a personalized approach. The quality of planning directly influences the feasibility of attaining rehabilitation objectives and has repercussions for various stakeholders within the healthcare institution, apart from the patients themselves. The article presents a comprehensive model depicting the life cycle of the cancer rehabilitation process. The article introduces a mathematical model that facilitates the automation of planning the progression of the rehabilitation process by utilizing predictive models. The techniques employed to model the system involve a fusion of approaches encompassing the modeling of physiological processes, statistical analysis, and machine learning. The suggested system enables the automation of cancer patient rehabilitation planning, mitigating the risk of complications, and optimizing the utilization of medical resources, thus minimizing the potential for conflicts of interest within the medical institution. Given that rehabilitation spans multiple stages and extends over a considerable duration, the authors employed Recurrent Neural Networks as predictive algorithms. In their pursuit of identifying the optimal approach for forecasting a set of pivotal benchmarks in rehabilitation assessment, the authors conducted a comparative analysis between conventional Recurrent Neural Networks and an Ensemble of Recurrent Neural Networks.