标题: | 经由支撑向量机进行多类别分类的线上学习来管理通讯网路流量 Online Learning by SVM for Multiclass Classification in Communication Network Traffic Management |
作者: | 苏建绮 卢鸿兴 Su, Jian-Chi Lu, Horng-Shing 统计学研究所 |
关键字: | 监督式学习;支撑向量机;线上学习;多类别分类;Supervised learning;Support vector machine;Online learning;Multiclass classification |
公开日期: | 2016 |
摘要: | 根据监督式学习里的支撑向量机方法,支撑向量机对于分类那些复杂的资料时是一个非常实用的方法。然而,当训练的资料集规模属于非常庞大的情况下,那么此方法在训练模型的过程中将会导致非常高的计算成本。线上学习的方法将会考虑妥善地处理这些关于记忆体的限制以及计算复杂度的相关问题。在此研究中,我们发展了一些经由线上学习的支撑向量机来处理关于多类别分类以及大数据资料的问题。我们将使用真实的通讯网路流量管理资料来评估这些方法的经验表现。 Supervised learning based on the method of support vector machine (SVM) is very useful for the classification of complex data. However, the computation cost is very high when the training dataset is massive. Online learning problems will need to handle the problems of memory limitation and computational complexity. In this study, the online learning methods by SVM for multiclass problems in massive data are developed. The empirical performance of these methods will be evaluated by real data in communication network traffic management. |
URI: | http://etd.lib.nctu.edu.tw/cdrfb3/record/nctu/#GT070352601 http://hdl.handle.net/11536/138360 |
显示于类别: | Thesis |