Dimensionality Reduction Effect Analysis of EEG Signals in Cross-Correlation Classifiers Performance

摘要: In this paper, it is reported a study conducted to verify whether the dimensionality reduction of electroencephalogram (EEG) segments can affect application performance machine learning (ML) methods. An experimental evaluation was performed in set 200 EEG segments, which piecewise aggregate approximation (PAA) method applied for 25 %, 50 and 75 % settings original segment length, generating three databases. Afterwards, cross-correlation (CC) these databases order extract features. Subsequently, classifiers were built using J48, 1NN, BP-MLP algorithms. These evaluated by confusion matrix method. The found that length increase or maintain ML methods, compared from with differentiate normal signals seizures.