说明：  非平衡数据集的分类问题经常出现在许多实际应用中.支持向量机在处理这一类问题时,整体分类性能比较低.为此,Veropoulos提出的采用不同惩罚系数的改进算法可以较好的解决此类问题.此外,可以利用序列最小优化算法简单快速的解决上述优化问题.
(Non-equilibrium data sets classification problems often appear in many practical applications. Support Vector Machine in dealing with this type of issue, the overall classification performance is relatively low. To this end, Veropoulos proposed punishment coefficient using different algorithms can improve the better solution such issues. In addition, can use the Sequential Minimal Optimization algorithm is simple and fast solution to the above optimization problem.)