Forming the System with the Functionality of Clinical Pharmacist for Personalized Treatment Strategy Searching

Autor:	Viktoria Iakymchuk, Maksym Suvorov, Oleksandr Matviichuk, Olena Nosovets, V.S. Pavlov, Ievgen Nastenko, Vitalii Babenko
Rok vydání:	2021
Předmět:	Decision support system Optimization problem business.industry Computer science Analytic hierarchy process Sample (statistics) Machine learning computer.software_genre Multi-objective optimization Task (project management) Random forest Genetic algorithm Artificial intelligence business computer
Zdroj:	Proceedings of Sixth International Congress on Information and Communication Technology ISBN: 9789811623769 ICICT (1)
DOI:	10.1007/978-981-16-2377-6_47
Popis:	Today, in Ukraine the task of creating an informative system for personalized treatment strategy searching is quite relevant. Its appearance will help to solve the long-burning issue of the absence of clinical pharmacists in medical institutions. These specialists are necessary because doctors cannot make an independent decision regarding the patient’s treatment strategy, since even their full compliance with the clinical protocols does not ensure a fully optimal human state. This work describes how the informative system can be formed, which will not only serve as a decision support system for doctors but also will effectively find the necessary solutions. The personalized treatment strategy searching is a multi-objective optimization problem since the final state of the patient is described by several indicators. To solve this kind of problem, the authors developed an approach using the principles of the genetic algorithm and the analytic hierarchy process. This approach was used in the practical task of finding personalized treatment strategies for patients with congenital heart defects to demonstrate the difference between the decision made by doctors in real life and the decision produced by the algorithm. Predictive models of indicators after treatment were obtained by the random forest classifier algorithm. Most of the models had 100% accuracy on the testing sample, which indicates the high efficiency of the used classification method. The promoted approach will be the foundation for the informative system developed in the future, and medical institutions can use it for any type of task regardless of the disease types.
Databáze:	OpenAIRE
Externí odkaz:	https://explore.openaire.eu/search/publication?articleId=doi_________::b27f09534bb9c6d2bacc73ff0aef0fbe https://doi.org/10.1007/978-981-16-2377-6_47 Zobrazit plný text záznamu