Data centering in feature space

  title={Data centering in feature space},
  author={Marina Meila},
This paper presents a family of methods for shifting the data in feature space to be used in conjunction with kernel machines. The shift can be performed using only kernel evaluations in input space. The methods are used to improve the numerical properties of kernel machines and to generate new families of string kernels. Experiments on real and artificial data show the beneficial effects of the centering methods and reveal further insights into the geometry of origin shifts in feature space. 
Highly Cited
This paper has 23 citations. REVIEW CITATIONS
17 Citations
14 References
Similar Papers


Publications referenced by this paper.

Similar Papers

Loading similar papers…