Sukhoi Daniil Vladimirovich (Perm State National Research University, Perm, Russia)
Polidanov Maxim Andreevich (research department specialist, assistant of the Department of Biomedical Disciplines, University «Reaviz», Saint Petersburg, Russia; Postgraduate student of the Department of Surgical Diseases, Medical University «Reaviz», Saratov, Russia)
Barulina Marina Aleksandrovna (Doctor of Physical and Mathematical Sciences, Professor, Professor of the Department of Natural Sciences, Medical University «Reaviz», Saratov, Russia; Doctor of Physical and Mathematical Sciences, Director of the Institute of Physics and Mathematics, Perm State National Research University, Perm, Russia; Doctor of Physical and Mathematical Sciences, Head of the Laboratory «Analysis and Synthesis of Dynamic Systems in Precision Mechanics», Chief Researcher, Saratov Scientific Center of RAS, Institute of Problems of Precision Mechanics and Control of RAS, Saratov, Russia)
Maslyakov Vladimir Vladimirovich (Doctor of Medicine Sciences, Professor, Professor of the Department of Mobilization Preparation of Public Health and Disaster Medicine, Saratov State Medical University named after V.I. Razumovsky of the Ministry of Health of the Russian Federation, Saratov, Russia; Doctor of Medicine Sciences, Professor, Professor of the Department of Surgical Diseases, Medical University «Reaviz», Saratov, Russia)
Parshin Alexey Vladimirovich (Candidate of Medical Sciences, Associate Professor, Associate Professor of the Department of Obstetrics and Gynecology of the Faculty of Medicine, Saratov State Medical University named after V.I. Razumovsky of the Ministry of Health of the Russian Federation, Saratov, Russia)
Volkov Kirill Andreevich (Saratov State Medical University named after V.I. Razumovsky of the Ministry of Health of the Russian Federation, Saratov, Russia)
| |
Early detection and prevention of complications remain a challenge due to individual patient characteristics and complex interaction of various risk factors. The purpose of the study. To investigate different approaches to develop a system for predicting complications arising from peritonitis using machine learning algorithms. Materials and methods of research. Data from 1558 patients hospitalized with the diagnosis of peritonitis were examined. The mean age was 44±8 years. To achieve the goal, a subtask was solved to find exactly those indicators that influenced the occurrence of complications. The constructed model showed good predictive ability in predicting the occurrence or non-occurrence of complications. In search of a reliable predictive model, we evaluated three most dissimilar and generally recognized as the most effective algorithms of classical machine learning: K-nearest neighbors (KNN), support vector machines (SVM), and gradient bousting on decision trees. Results. So, we investigated different approaches for developing a system for predicting the probability of peritonitis complication probability by artificial intelligence methods. For the problem at hand, the most effective algorithm was found to be KNN, which was tuned using cross-validation and grid-based hyperparameter search. Conclusion. As future work, the model should continue to be monitored and updated to reflect changing trends in data and medical practice. The study demonstrates the potential of using machine learning algorithms to improve the prediction of complications of peritonitis, thereby helping to improve the quality of patient care and treatment.
Keywords:Peritonitis; early detection of complications; prediction of complications; machine learning
|
|
| |
|
Read the full article …
|
Citation link:
Sukhoi D. V., Polidanov M. A., Barulina M. A., Maslyakov V. V., Parshin A. V., Volkov K. A. PROBLEMS OF CREATING SYSTEMS FOR PREDICTING COMPLICATIONS IN PERITONITIS BASED ON MACHINE LEARNING METHODS // Современная наука: актуальные проблемы теории и практики. Серия: Естественные и Технические Науки. -2024. -№07. -С. 139-146 DOI 10.37882/2223-2966.2024.7.36 |
|
|
