摘要
脑神经信号属于非平稳随机信号,是一类难以进行分类识别的信号。为改进这类信号的分类效果,提出了将灰关联理论应用于脑神经信号的分类。首先介绍了灰关联理论和方法,在此基础上,建立了脑神经信号灰色模型(Grey Model)——GM(1,1)模型,估计出每一个模型参数a和b,将其中模型参数b作为特征值用于灰关联分析,得到第1次分类结果,然后在认真分析第1次分类结果的基础上,进一步给出二次分类方法。通过二次分类,实现了对脑神经信号的有效分类识别,其分类正确率高达88%。结果表明,将灰关联技术用于非平稳随机信号的分类与识别是可行而有效的,有很好的应用前景。
Aim. Among Nonstationary Randomness Signals（NRSs）, brain neural signals are particularly difficult to classify. We now propose applying grey correlation theory to improving considerably the classification of brain neural signals. In the full paper, we explain in detail improving the classification of brain neural signals and analyzing the results of improvement ; in this abstract, we just add some pertinent remarks to listing the three topics of explanation： （1）grey correlation analysis, （2）the numerical example of classifying brain neural signals with grey correlation analysis and （3）the results of analysis： in topic 2, we establish the GM（1,1） models of brain neural signals and obtain eigenvalues a and b for each GM （1, 1） model; also in topic 2, we use all the eigenvalue b＇s to perform grey correlation analysis, thus obtaining the first-time classification results as summarized in Table 2 in the full paper; in topic 3, we, on the basis of careful analysis, explain how to utilize the results of Table 2 to improve classification, thus obtaining the second-time and final classification results as summarized in Table 5 or Table 6 in the full paper; also in topic 3, Table 6 shows that the improved correct classification rate reaches as high as 88%.
出处
《西北工业大学学报》
EI
CAS
CSCD
北大核心
2007年第1期122-125,共4页
Journal of Northwestern Polytechnical University
基金
国家自然科学基金(30470459)
西北工业大学科技创新基金(M450212)资助
关键词
灰关联
非平稳随机信号
脑神经信号
特征值
正确分类率
grey correlation analysis, Nonstationary Randomness Signal （NRS）, brain neural signal, eigenvalue, correct classification rate
