23 августа 2021

Hybrid Predictive Modelling for Finding Optimal Multipurpose Multicomponent Therapy

161

Pavlovskii V.V., Derevitskii I.V., Kovalchuk S.V.

International Conference on Computational Science

Абстракт: This study presents a new hybrid approach to predictive modelling of disease dynamics for finding optimal therapy. We use existing methods, such as expert-based modelling methods, models of system dynamics and ML methods in compositions together with our proposed modelling methods for simulating treatment process and predicting treatment outcomes depending on the different therapy types. Treatment outcomes include a set of treatment-goal values, therapy types include a combination of drugs and treatment procedures. Personal therapy recommendation by this approach is optimal in terms of achieving the best treatment multipurpose outcomes. We use this approach in the task of creating a practical tool for finding optimal therapy for T2DM disease. The proposed tool was validated using surveys of experts, clinical recommendations [1], and classic metrics for predictive task. All these validations have shown that the proposed tool is high-quality, interpretable and usability, therefore it can be used as part of the Decision Support System for medical specialists who work with T2DM patients.

Pavlovskii V.V., Derevitskii I.V., Kovalchuk S.V. (2021) Hybrid Predictive Modelling for Finding Optimal Multipurpose Multicomponent Therapy. In: Paszynski M., Kranzlmüller D., Krzhizhanovskaya V.V., Dongarra J.J., Sloot P.M. (eds) Computational Science – ICCS 2021. ICCS 2021. Lecture Notes in Computer Science, vol 12744. Springer, Cham (WoS/Scopus Q3/4). https://doi.org/10.1007/978-3-030-77967-2_40

Поделиться

Подпишитесь на нашу рассылку

Хотите получать интересную и полезную информацию о цифровом здравоохранении и искусственном интеллекте для медицины?
Включайтесь в нашу рассылку!

Присоединяйтесь

Наши группы в соц сетях