结合类标签关联度的有序核判别回归学习  

Kernel Discriminant Learning for Ordinal Regression Using Label Membership

作　　者：李亚克 田青[1] 高航[1] 

机构地区：[1]南京航空航天大学计算机科学与技术学院,南京210016

出　　处：《数据采集与处理》2016年第3期532-540,共9页Journal of Data Acquisition and Processing

摘　　要：有序离散类标号通常由原始连续标号按一定规则映射得到,因此它们彼此间是存在关联信息的,现有有序回归方法对此类关联信息的考虑仍然较少。首先提出一类有序标号间关联度的量化表示,进而将其与典型有序回归方法(Kernel discriminant learning for ordinal regression,KDLOR)相结合,设计出了一种结合类标号关联度的有序核判别回归学习方法(Kernel discriminant learning for ordinal regression using label membership,LM-KDLOR),最后通过在多个标准有序回归数据集上的对比实验验证了所提方法的有效性。Abstract： The ordinal discrete labels are usually obtained from continuous labels, and these regressor sel- dom use the mutual membership information between ordinal discrete labels, which can be further im- proved. Therefore, quantitive representation is characterized for the membership information, and then a kernel diseriminant learning for ordinal regression using label membership （LM-KDLOR） is established by combining the representation with typical off-the-shelf KDLOR. Experimental results with the stand- ard ordinal regression data sets verify the effectiveness of the proposed strategy.

关 键 词：有序回归 核判别学习有序回归 类标签关联度 核方法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]