The clustering performance of sdfcm is researched ; the advantages of sdfcm compared with traditional methods , fcm and other artificial intelligent methods are pointed out . to sum up , the main creative results include : 1 综合上述研究内容，本文的主要创新点包括： l针对fcm采用欧氏距离造成的分类错误，提出了反映制造单元构建问题本质特征的距离函数。
We analyze the classifying results based on the fuzzy text classifying , think the wrong classifying results can be divided into two styles , and we propose a subordinative degree update algorithm aim at the two instances . combined the nizzy semantic relationship classifying algorithm , we propose the gradual classifier construction algorithm through checkouting and correcting the wrong results constantly with the update formula . 在模糊文本分类的基础上，对分类结果进行了分析，将分类错误归结为两种类型，并针对这两种情况提出了隶属度更新算法，结合模糊语义关联度的分类算法提出了运用更新公式不断对分类结果进行校验纠错进而逐渐地构造分类器的算法。
Normal behavior and anomaly are distinguished on the basis of observed datum such as network flows and audit records of host . when a training sample set is unlabelled and unbalanced , attack detection is treated as outlier detection or density estimation of samples and one - class svm of hypersphere can be utilized to solve it . when a training sample set is labelled and unbalanced so that the class with small size will reach a much high error rate of classification , a weighted svm algorithm , i 针对训练样本是未标定的不均衡数据集的情况,把攻击检测问题视为一个孤立点发现或样本密度估计问题,采用了超球面上的one - classsvm算法来处理这类问题;针对有标定的不均衡数据集对于数目较少的那类样本分类错误率较高的情况,引入了加权svm算法-双v - svm算法来进行异常检测;进一步,基于1998darpa入侵检测评估数据源,把两分类svm算法推广至多分类svm算法,并做了多分类svm算法性能比较实验。
