本文在参数关系的独立与联合相关中取适当的折衷,考虑了参数量级之间的相关性,用上下文方法对脑电图数据进行分类识别,取得比一般统计方法及模糊数学方法好的结果.脑电图识别是根据临床测试获得的数据,给出正常或异常程度的诊断.根据经验,对脑电波参数抽取5项测试指标:频率F,3量级;振幅A,4量级;调幅周期TT,3量级;调幅间歇JT,2量级;H.V.试验HV,2量级,分别在头的左部和右部的4部位,即额额(F-T)、颞枕(T-O)、额顶(F-P)和顶枕(P-O),对这5项测试指标进行测试.每个测试点的每个测试指标参数由其范围的上下限中两个或一个数据表示,共得136个测试数据, In this paper, we take the appropriate trade-off between the independent and the joint correlation of parameter relations, considering the correlation between the order of parameters, using the contextual method to classify the EEG data and get the better results than the general statistical methods and fuzzy mathematics methods EEG recognition is based on the data obtained from clinical tests and gives a diagnosis of normal or abnormal levels.Experimental results show that the EEG parameters are extracted from five test parameters: frequency F, magnitude 3, amplitude A, magnitude 4, amplitude modulation Period TT, 3 orders of magnitude; amplitude modulation intermittent JT, 2 orders of magnitude; HV test HV, 2 orders of magnitude, respectively, in the left and right head 4 parts, Top (FP) and Top Pill (PO) to test these five test indexes.Each test index parameter of each test point is represented by two or one data in the upper and lower limits of its range, a total of 136 test data ,