부산대학교 도서관

Epilepsy is a non-contagious condition of the brain which affects individualsof all age groups, with approximately 60% of people worldwide suffering from it. The unpredictable nature of epileptic seizures necessitates accurate diagnosis for timely intervention. Electroencephalography (EEG) is a pivotal tool in epilepsy diagnosis, as it captures transient changes in brain electrical activity. This paper presents a comparative analysis of various machine learning algorithms applied to EEG data for classifying different seizure types. Our approach focuses on improving performance through data augmentation and preprocessing techniques. We explore the efficacy of algorithms such as Decision Trees, Gaussian Naive Bayes, Multi-Layer Perceptrons. K-Nearest Neighbors, Random Forest, and Support Vector Classifier. The results highlight the superior performance of Multi-Layer Perceptron, achieving an accuracy of 96%, surpassing other algorithms. This research underscores the possibility of machine learning in enhancing epilepsy diagnosis and seizure prediction, contributing to improved patient care.