Analysis of brain thermometric data obtained by microwave radiothermometry

Бесплатный доступ

This paper discusses the effectiveness of using the method of microwave radiothermometry in examinations of brain diseases, namely the state of disordered consciousness. In contrast to most methods of examinations by this method, the measurements of the brain were carried out in only 2 different frontal areas for 2 days with a frequency of 4 hours. Therefore, the aim of the study was to identify the effectiveness of a diagnostic model based on the dynamics of temperature changes. The work showed that in healthy patients there is a circadian rhythm: during the day the temperature rises, at night it decreases. At the same time, such dynamics is not observed in patients with disordered consciousness. Based on this knowledge, a conceptual and mathematical modelwere proposed. The first of them describes the characteristic features of healthy and sick patients. The second one quantifies these features. The constructed mathematical model was tested in the classification problem. The Naive Bayes classifier was used as a classifier. As a result of computational experiments, it was shown that for 500 iterations the classifier made a mistake on only1 sick patient and 5 healthy ones. Thus, the effectiveness of the method of microwave radiothermometry in the task of examining patients with disordered consciousness was shown.

Еще

Microwave radiothermometry, machine learning, data mining, circadian rhythm, classification algorithm

Короткий адрес: https://sciup.org/149143813

IDR: 149143813   |   DOI: 10.15688/mpcm.jvolsu.2023.2.3

Статья научная