INDEFINITE KERNEL NETWORK WITH DEPENDENT SAMPLING
INDEFINITE KERNEL NETWORK WITH DEPENDENT SAMPLING
复制标题
DOI：
10.1142/s0219530513500206
复制
发表时间：
2013-09
影响因子：
2.2
通讯作者：
Hongwei Sun;Qiang Wu
中科院分区：
文献类型：
--
作者：
Hongwei Sun;Qiang Wu
研究了系数正则化和相依采样的不定核网络的渐近性质。所研究的框架不同于经典的内核学习。核函数不要求正定性，允许样本弱相关，相关性由强混合条件度量。通过[27]中引入的一种新的核分解技术，利用两个再生核Hilbert空间及其相应的核积分算子来刻画假设函数类的性质和可学习性。推导了容量无关的误差界和学习率。
We study the asymptotical properties of indefinite kernel network with coefficient regularization and dependent sampling. The framework under investigation is different from classical kernel learning. Positive definiteness is not required by the kernel function and the samples are allowed to be weakly dependent with the dependence measured by a strong mixing condition. By a new kernel decomposition technique introduced in [27], two reproducing kernel Hilbert spaces and their associated kernel integral operators are used to characterize the properties and learnability of the hypothesis function class. Capacity independent error bounds and learning rates are deduced.