How to extract effective discriminative information from overlapping(outlier)samples is still open.

• 如何从重叠(离群)样本中抽取有效的分类特征没有得到有效的解决。