论文部分内容阅读
针对现有单类分类器对目标数据先验信息考虑的不足,在结构单类支持向量机(structured one-class supportvector machine,SOCSVM)中嵌入局部密度信息,提出局部密度嵌入的结构单类支持向量机(SOCSVM with local den-sity embedding ldSOCSVM)。借助K近邻(K-nearest neighbor,KNN)揭示目标数据局部密度,并进一步诱导出权重因子作用于样本点。该算法充分利用目标数据的全局信息及局部密度信息,从而提高分类器的泛化能力。UCI数据集上的实验结果验证了ldSOCSVM的有效性。
In order to overcome the shortcomings of existing single classifiers in priori information of target data, local density information is embedded in a structured one-class support vector machine (SOCSVM), and a structure-based single-class support vector SOCSVM with local den-sity embedding ldSOCSVM. The K-nearest neighbor (KNN) is used to reveal the local density of the target data and further induce the weight factor to act on the sample points. The algorithm makes full use of the global information and local density information of the target data to improve the generalization ability of the classifier. The experimental results on the UCI dataset validate the validity of ldSOCSVM.