Institutional Repository of Vilnius University
Mašininio mokymo taikymai EEG: lyties nustatymas /
Abstract
dc:descriptionMachine Learning (ML) approaches have been integrated in all areas of a modern EEG analysis pipeline. From signal processing with Independent Component Analysis (ICA), to microstate analysis using K-Means and subject classification using tree-based and deep learning ML models. This thesis walks through the extraction of features from EEG data, acknowledging current research that supports each step. Additionally, we have demonstrated the application of reviewed methods. On an open-source set of real subjects $(N=165)$, using an eXtreme Gradient Boosting (XGBoost) model to classify sex with a $76.5\%$ Receiver Operating Characteristic-Area Under the Curve (ROC-AUC) accuracy. SHAP value analysis revealed the most significant features in determining sex from EEG data. These features coincide with current research, indicating the strong predictive power of microstate, complexity and frequency power metrics.
Degree
thesis:*- Grantor dc:publisher
- Institutional Repository of Vilnius University
- Year dc:date
- 2026
Author and committee
dc:creator, dc:contributor.*- Author dc:creator
-
- Rastenis, Vilius,
Rights
dc:rights- Statement dc:rights
-
- info:eu-repo/semantics/openAccess
- Language dc:language
- lit
Identifiers
dc:identifier.*- Repository record dc:identifier
- https://repository.vu.lt/VU:ELABAETD308327012&prefLang=en_US