Robust ordinal regression induced by l(p) -centroid
Robust ordinal regression induced by l(p) -centroid
复制标题
由 lp 质心诱导的稳健序数回归
DOI：
10.1016/j.neucom.2018.06.041
复制
发表时间：
2018
期刊：
影响因子：
6
通讯作者：
Yin Hujun
中科院分区：
文献类型：
--
作者：
Tian Qing;Zhang Wenqiang;Wang Liping;Chen Songcan;Yin Hujun
有序回归（OR）是机器学习领域的一个重要研究课题，因其广泛的应用而受到广泛关注。到目前为止，已经提出了各种方法来执行或，其中类中心诱导阈值方法（如KDLOR和莫尔）由于其简单和有前途的性能而受到更多的关注。类中心诱导的OR通常计算具有类中心的有序阈值，其通常从12范数导出。不幸的是，以这种方式，当数据被离群值破坏时，类均值可能有偏差（即，非i.i.d.噪声），使得所得到的OR准确度将劣化。受lp-范数在抗噪声应用中的成功启发，本文提出了一种新的由lp-范数导出的类质心（简称aslp-质心）来克服上述缺点，并给出了计算lp-质心的优化算法和相应的收敛性分析。为了评估p-重心在OR上下文中对抗噪音的有效性，我们将p-重心与两个代表性的类中心诱导的OR（即基于判别学习的OR和基于流形学习的OR）联合收割机起来。最后，在合成数据集和真实数据集上的大量OR实验证明了所提方法的有效性和优越性。
Ordinal regression (OR) is an important research topic in machine learning and has attracted extensive attention due to its wide applications. So far, a variety of methods have been proposed to perform OR, in which the class-center-induced threshold methods (like KDLOR and MOR) have received more attention, for their simplicity and promising performance. The class-center-induced ORs typically calculate the ordinal thresholds with class centers, which are typically derived from thel2-norm. Unfortunately, in such a way, the class means may be biased when the data is corrupted with outliers (i.e., non-i.i.d. noises) such that the resulting OR accuracy will be deteriorated. Motivated by the success oflp-norm in applications against noises, in this paper we propose a novel type of class centroid derived from thelp-norm (coined aslp-centroid) to overcome the drawbacks above, and provide an optimization algorithm and corresponding convergence analysis for computing thelp-centroid. To evaluate the effectiveness oflp-centroid in OR context against noises, we then combine thelp-centroid with two representative class-center-induced ORs, namely discriminant learning based and manifold learning based ORs. Finally, extensive OR experiments on synthetic and real-world datasets demonstrate the effectiveness and superiority of the proposed methods to related existing methods.