Fast Dynamic Time Warping Feature Extraction for EEG Signal Classification

Abstract

In this work the fast algorithm Dynamic Time Warp (FDTW) is used as a method of feature extraction for 18 sets of EEG records. Each set contains 150 events of stimulation designed to study the semantic relationship between pairs of nouns of concrete objects such as "HORSE - SHEEP" and "SWING - MELON" and how this relationship activity is reflected in EEG signals. Based on these latter, different classifiers were trained in order to associate a set of signals to a previously learned human answer, pertaining to two classes: semantically related, or not semantically related. The results of classification accuracy were evaluated comparing with other 3 methods of feature extraction, and using 5 different classification algorithms. In all cases, classification accuracy was benefited from using FDTW instead of LPC, PCA or ICA for feature extraction.