International Conference PhysicA.SPb/2023

The application of machine learning methods in searching for statistical patterns for diagnosing obsessive-compulsive disorder

V. A. Yunusov, S. A. Demin

Kazan (Volga Region) Federal University, Kazan, Tatarstan, Russia

Abstract: One of the urgent tasks of modern data sciences is defining diagnostic criteria for mental disorders. This task is complicated by the existence of many biophysical parameters, some of which may be redundant. In this paper, we apply techniques for selecting features necessary to diagnose the obsessive-compulsive disorder. With the help of machine learning methods, the classification problem was solved for the initial set of features at the first stage of work; at the second stage, subsets of the most significant diagnostic features were selected for volunteers exhibiting significant symptoms of this disorder as well as for representatives of the reference group.

Keywords: living systems, obsessive-compulsive disorder, biomedical data, machine learning methods, feature attribution.

Received: 05.05.2023
Revised: 29.06.2023
Accepted: 30.10.2023

DOI: 10.61011/PJTF.2023.23.56840.13A

Bibliographic databases:

•