Recently, kernel-based non-parametric models e.g., [ 15- 18] have been proposed.
Only recently, kernel-based methods have also been used in association studies (K wee et al. 2008; Y ang et al. 2008) and QTL mapping for complex traits (Z ou et al. 2010), which demonstrates their great potential and flexibility.
By considering the distinct features of MHC-II peptide binding prediction problem, MHC2SK differs significantly from the recently developed kernel based method, GS (Generic String) kernel, in the way of computing similarities.
Recently, reproducing kernel methods (RKMs) were used to solving a variety of BVPs [13 24].
This paper is a follow-up study of the recently introduced kernel least-mean-square algorithm (KLMS).
A recently developed Kernel Density Estimation-based model that can adequately represent multimodal wind data is employed to characterize the wind distribution.
We introduce a family of kernels capturing the similarity of fragmentation trees, and combine these kernels using recently proposed multiple kernel learning approaches.
Recently, several kernel-based supervised network inference methods have been developed (Vert and Yamanishi, 2005; Yamanishi et al., 2004), but they are limited to interactions between homogeneous molecules (e.g. protein protein interactions) with a simple graph representation.
In terms of runtime on large graphs, these kernels outperform other kernels, including the recently developed random walk kernels [ 8] and graphlet kernels [ 26].
The type of features considered are those developed for a recently proposed graph kernel called Neighborhood Subgraph Pairwise Distance Kernel (NSPDK) (Costa and Grave, 2010) and used for the efficient clustering of ncRNA molecular graphs in Heyne et al. (2012).
