Predicting Seizure Onset with Intracranial Electroencephalogram (EEG) Data - Project Report

Using over 3,000 training examples, we investigated multiple feature extraction and classification algorithms including discrete wavelet transform (DWT), short-time Fourier transform (STFT), principal component analysis (PCA), k-Nearest Neighbors, logistic regression, and support vector machines. Using STFT, PCA, and logistic regression we developed a feature extraction and classification algorithm to classify EEG signals as preictal or interictal with an Area Under Curve (AUC) score of .75.