P300信号分类的多样本融合支持向量机算法

脑电信号是大脑受到自发或诱发刺激所产生的一种变化的脑电活动，利用结合了多样本融合思想的支持向量机（SVM）算法，在不同受试者的多样本数据上对诱发脑电信号中的P300信号进行了分类识别。首先对实验数据进行预处理和特征提取，然后利用SVM算法训练分类模型，最后通过融合多个样本的预测结果对测试数据的P300信号进行识别。结果表明，相比单样本SVM算法该方法对检测数据有较高的分类准确率，能够成为P300脑电信号预测的较好方法，具有应用前景。

Support Vector Machine Algorithm with Multi-Sample Fusion for P300 Signal Classification

Electroencephalogram(EGG) signal is a type of altered EEG activity produced by spontaneous or evoked stimulus of the brain. A support vector machine (SVM) algorithm with multi-sample fusion was applied for the classification of the P300 in the multi-sample evoked EEG signals from different subjects. First, the experimental data were preprocessed and the features were extracted, then the classification models were trained by SVM, and finally the prediction results of different samples were integrated in identifying the P300 signal of the test data. The results show that the proposed method has higher?classified?accuracy than the single-sample SVM algorithm, it can become a better method for P300 EEG signal prediction with application prospects.