Sequential Pattern Learning via Kernel Alignment
As a branch of data analysis, pattern alignment has received much attentions in recent years. More specifically, it learns to find intrinsic bridge between different domains and make data handling be transferrable for efficient recognition. In this work, an unsupervised feature learning method is proposed to meet demand on pattern alignment. Compared with existing methods, more efficiency can be reached owing to scalable learning, which is competent to tackle large-scale data for kernel alignment. Experimental results show proposed method can give comparable performance among the state-of-The-Art methods.