论文部分内容阅读
针对运动想象(MI)脑电信号识别精度低的实际问题,提出了一种结合免疫优化算法和决策机制的堆叠降噪自编码机网络(ISDAE).ISDAE模型通过多层DAE对MI脑电信号进行分层提取最优特征向量,再通过最后一层NN网络,对所得特征向量进行识别;同时,添加决策机制,并结合免疫优化算法对模型进行参数寻优,最终得到识别准确率更高的ISDAE脑电信号识别模型.实验结果表明,本文提出的ISDAE模型对粗糙的脑电数据具有强大的特征学习能力和较高的MI脑电信号识别率,为MI脑电信号的识别提供了一个有效的方法.
Aiming at the practical problem of low recognition accuracy of motor imagery (MI), a stack noise reduction self-encoder network (ISDAE) based on immune optimization algorithm and decision-making mechanism is proposed.ISDAE model uses multi-layer DAE to measure MI brain electrical signal Then the optimal eigenvector is extracted by layer, and then the final eigenvector is identified through the last NN network. At the same time, the decision-making mechanism is added, and the immune optimization algorithm is used to optimize the parameters of the model. Finally, the identification accuracy is higher ISDAE EEG recognition model.The experimental results show that the proposed ISDAE model has strong feature learning ability and high MI EEG recognition rate for rough EEG data and provides an effective method for the identification of MI EEG signals Methods.