论文部分内容阅读
在基于支持向量机的多分类算法中,一对一算法表现出较好的性能.然而此算法却存在不可分区域,落入该区域的样本不能有效被识别,因此影响了一对一算法的性能.为解决这个难题,提出交互迭代一对一分类算法,同时给出算法的有效性分析和计算复杂度证明.为了验证该算法解决不可分区域的能力,我们选用 UCI 数据集来做对比实验.实验结果显示,本文算法不但可以较成功解决不可分区域问题而且表现出比其它算法更好的性能.
In the multi-classification algorithm based on SVM, the one-to-one algorithm shows better performance, however, this algorithm has inseparable regions and the samples falling into this region can not be effectively identified, thus affecting the performance of one-to-one algorithm In order to solve this problem, a one-to-one classification algorithm based on iterative iteration is proposed, and the validity of the algorithm and the proof of computational complexity are also given.In order to verify the capability of the algorithm to solve the problem of inseparable regions, we use UCI dataset to do comparative experiments. The results show that the proposed algorithm can not only solve the problem of indivisible region more successfully, but also show better performance than other algorithms.