2D Numerical Dataset for Microwave SVM-Based Brain Stroke Classification

In this study, we investigated microwave stroke detection and classification using machine learning algorithms. To obtain large datasets with high data variability, we utilized two distinct 2D numerical models. Next, we employed PCA to reduce the data dimensions and evaluated classification performa...

Full description

Saved in:
Bibliographic Details
Published in2023 Photonics & Electromagnetics Research Symposium (PIERS) pp. 1705 - 1711
Main Authors Pokorny, Tomas, Fiser, Ondrej, Drizdal, Tomas, Vrba, Jan
Format Conference Proceeding
LanguageEnglish
Published IEEE 03.07.2023
Subjects
Electromagnetics
Filling
Machine learning algorithms
Microwave antennas
Numerical models
Numerical simulation
Training
Online AccessGet full text

Cover

More Information
Summary:In this study, we investigated microwave stroke detection and classification using machine learning algorithms. To obtain large datasets with high data variability, we utilized two distinct 2D numerical models. Next, we employed PCA to reduce the data dimensions and evaluated classification performance of six different machine learning algorithms. Additionally, we investigated the impact of the way how the matching medium is placed in front of the antennas, which enhanced the variability of the principal components. Despite this improvement, we observed only a slight increase in the accuracy of stroke classification.
ISSN:2831-5804
DOI:10.1109/PIERS59004.2023.10221426