集群SVM大规模数据分类算法

Ensemble SVM classification for large scale data sets

为了提高SVM在大规模数据集上的训练效率和检测精度,对训练数据预处理后进行无监督聚类,通过一定规则选取对训练SVM有用的样本向量,并结合改进的AdaBoost算法来增强SVM在大规模数据的分类和泛化能力,最后通过Kdd Cup 99数据进行实验验证算法性能.

In order to improve the efficiency and precision of the SVM that training on large scale data sets, after preprocessing the data,we run the unsupervised clustering which holds at certain rules by which selecting a sample of training vectors that are useful for SVM, Then incorporate the enhanced AdaBoost algorithm to improve the SVM ability for classification and generalization,Finally we use dataset Kdd Cup 99 to verified performance of the algorithm.