论文部分内容阅读
提出了基于最优超平面与支持向量机思想的最大间隔聚类算法。该方法借鉴了最优超平面思想和用核函数非线性映射构造支持向量机的思想。通过构造一个二次规划问题 ,得到了使分类后两类间距最大的聚类方法 ,并且借助非线性核函数将该方法推广到非线性情况。仿真试验表明 :该方法可以较好地解决很多非监督分类问题 ,得到的结果基本不受数据分布形状的影响
A maximum interval clustering algorithm based on the idea of optimal hyperplane and support vector machine is proposed. The method draws on the idea of optimal hyperplane and construction of support vector machines by kernel mapping nonlinear mapping. By constructing a quadratic programming problem, the clustering method that maximizes the separation between the two classes is obtained. The nonlinear kernel function is used to extend the method to non-linear case. The simulation results show that this method can solve many unsupervised classification problems well, and the results are basically independent of the shape of data distribution