Title :
A study of data reduction for P300 speller system
Author :
Rittikun, Kittisak ; Boonpramuk, Panuthat ; Prechaprapranwong, Prapong
Author_Institution :
Dept. of Control Syst. & Instrum. Eng., King Mongkut´s Univ. Thonburi, Bangkok, Thailand
Abstract :
The aim of this research is to reduce data for a P300 spelling system by using downsampling algorithm to reduce the sampling rate of the system. The comparison between this method and standard algorithms such as Discrete Wavelet Transforms (DWT) and Principal Components Analysis (PCA) is also discussed in this paper. While downsampling algorithm is used to reduce sampled data, Ensemble of Support Vector Machine (ESVM) is used to model and to classify the reduced data in order to predict characters. The experimental results show that the accuracy of downsampling algorithm, no-data-reduction algorithm, DWT and PCA are 97.5%, 93.5%, 96.0%, and 79.5% respectively.
Keywords :
brain-computer interfaces; discrete wavelet transforms; electroencephalography; learning (artificial intelligence); medical signal processing; principal component analysis; sampling methods; signal classification; support vector machines; DWT; ESVM; P300 speller system; PCA; data reduction; discrete wavelet transforms; downsampling algorithm; ensemble of support vector machine; no-data-reduction algorithm; principal components analysis; reduced data classification; sampling rate reduction; Accuracy; Algorithm design and analysis; Discrete wavelet transforms; Information filters; Principal component analysis; Support vector machines; P300 speller; amyotrophic lateral sclerosis; discrete wavelet transforms; ensemble of classifiers; support vector machine;
Conference_Titel :
Electrical Engineering/Electronics, Computer, Telecommunications and Information Technology (ECTI-CON), 2014 11th International Conference on
Conference_Location :
Nakhon Ratchasima
DOI :
10.1109/ECTICon.2014.6839866
