Offline Detection of P300 in BCI Speller Systems

摘要: The paper presents a framework for offline analysis of P300 speller system using seeded k-means based ensemble SVM. Due to the use small-datasets training classifier, performance deteriorates. Proposed emphases on semi-supervised clustering approach SVM classifier with large amount data. normalized mutual information (NMI) has used cluster validation that gives maximum 88 clusters 10 fold cross-validation dataset NMI approx equals 1. proposed applied EEG data acquired from two subjects and provided by Wadsworth center brain-computer interface (BCI) competition III. experimental results show increase in SNR value obtain better accuracy than linear, polynomial or rbf kernel