【摘要】 最小二乘支持向量机采用最小二乘线性系统代替传统的支持向量即采用二次规划方法解决模式识别问题,能够有效地减少计算的复杂性。但最小二乘支持向量机失去了对支持向量的稀疏性。文中提出了一种基于边界近邻的最小二乘支持向量机,采用寻找边界近邻的方法对训练样本进行修剪,以减少了支持向量的数目。将边界近邻最小二乘支持向量机用来解决由1-a-r(one-against-rest)方法构造的支持向量机分类问题,有效地克服了用1-a-r(one-against-rest)方法构造的支持向量机分类器训练速度慢、计算资源需求比较大、存在拒分区域等缺点。实验结果表明,采用边界近邻最小二乘支持向量机分类器,识别精度和识别速度都得到了提高。更多还原

【Abstract】 Least squares support vector machines can reduce the high computational complexity.Duo to equality type constraints in the formulation,the solution follows from solving a set of linear equations,instead of quadratic programming for classical SVM.But unfortunately a very attractive feature of SVM,namely its sparseness,was lost.A new least squares support vector machines based on boundary nearest was proposed,which reduced the number of support vector by using boundary nearest methods pruning the training sample.Meanwhile,uses least square support vector machine for solving multiclass problems constructed by 1-a-r(one-against-rest) method,which overcame the demerits of low training speed,high computational requirements and existing reject area in 1-a-r(one-against-rest) method effectively.Experimental result shows that the performance of accuracy and speed of classifiers are improved after using least squares support vector machines based on boundary nearest.更多还原