半监督支持向量机综述.PDFVIP

半监督支持向量机综述 年素磊 (南京大学 计算机科学与技术系, 南京 210093) Semi-supervised Support Vector Machines Survey * Su-Lei Nian (Department of Computer Science and Technology, Nanjing University, Nanjing 210093, China) Abstract : Support vector machine as a machine learning method aimed at small samples is developed by Vapnik and others in the basis of statistical learning theory. It has been widely researched and applied for its advantage of strong generalization ability and convenient for high dimension data operation. Though standard support vector machine based on supervised learning can resolve many actual problems effectively, it have to label mass unlabeled data in order to get enough training samples. It makes these methods costly and low efficiency . So semi-supervised support vector machines are proposed according to the actual requirement. Over the past decade, many improved versions are proposed to solve every different specific problem of semi-supervised support vector machines, for example: to the inefficiency issue, meanS3VM is introduced; to performance degenerate problem, S4VM is proposed; and to cost sensitive issue, CS4VM emerges. Key words: semi-supervised support vector machine; statistical learning theory; inefficiency; performance degenerate; cost sensitive 摘 要: 支持向量机是Vapnik 等在统计学习理论基础上发展起来的针对小样本的机器学习方法。该方法由于 具有较强的泛化能力、方便对高维数据操作而得到了日益广泛的研究和应用。标准的支持向量机是基于监督 学习的，虽然能有效地解决各种实际问题，但需要手工对大量样本进行标记以获取足够的训练样本，代价高， 效率低。因此根据实际需要研究人员又提出了半监督支持向量机。在过去的十年中，为了解决半监督支持向 量机某一方面的特定问题，出现了很多改进版本，如针对半监督支持向量机效率低下的问题，提出了 meanS3VM 算法；针对利用无标记数据时会产生性能下降的问题，提出了S4VM 算法；针对代价敏感的问题， 提出了CS4VM 算法。 关键词: 半监督支持向量机;统计学习理论;低效;性能下降;代价敏感 中图法分类号: TP301 文献标识码: A 1 引言 随着计算机在日常生活的广泛应用，各行各业都积累了大量的数据。如何快速、准确、方便地从海量的 信息库中获取感兴趣、满足需要的信息，已成为人们亟待解决的问题。在各种复杂应用背景条件下，人工方  作者简介: 年素磊 南京大学计算机系 2011 研究生 2 式不可能处理如此庞大的信息，机器学习方法就显得尤为重要。机器学习方法是从观测样